Multi-objective network optimization method for plant community in intelligent logistics edge equipment

Abstract

A plant community multi-objective network optimization method in an intelligent logistics edge device completes intelligent logistics network multi-objective optimization by simulating behaviors of sowing, growing, flowering and fruiting of a plant community on the intelligent logistics edge device, and comprises the following steps: step 1, solving intelligent logistics network planning optimization by the plant community; step 2, optimizing a plant community solution intelligent logistics plan; step 3, performing optimization on plant community solution intelligent logistics execution; and step 4, optimization of the intelligent logistics reverse network for solving the plant community. According to the method, multi-objective optimization of the intelligent logistics network can be efficiently completed on the edge device.

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and in particular to logistics and artificial intelligence technology, and specifically to a multi-objective network optimization method for plant communities in smart logistics edge devices. Background Art

[0002] With the advancement of science and technology and the economy, advanced computer technology, automatic control technology, and big data technology have begun to be widely used in the field of logistics, and logistics services have begun to show a trend of intelligence and automation. Smart logistics uses a large number of smart express cabinets, drones, unmanned vehicles, electric vehicles, and satellite positioning data to provide precise services for enterprises and individuals. Smart logistics will become the mainstream of the logistics industry in the future. The drone delivery route is in the air, basically ignoring the impact of complex terrain. In areas with small land areas, high population density, high building density, and complex traffic conditions, drones can bypass obstacles and deliver packages directly to the destination.

[0003] Since 2016, the scale of China's express locker market has been growing at a high rate of more than 30% per year, and the locker rate has continued to grow at a rate of more than 25%, which played an important role in delivering goods to quarantined people during the three-year epidemic. Drone logistics can directly separate the mailing business and the delivery business. The goods mailing business is handed over to the express delivery station, while the drone is mainly responsible for the delivery of goods, which greatly reduces the manual workload of the station and improves the customer experience. In 2012, SF Express proposed the concept of drone logistics. In September 2013, SF Express conducted its first drone test in Dongguan City, Guangdong Province. In March 2018, SF Express obtained the first domestic drone aviation operation license. In August 2020, SF Express conducted its first large-scale drone cargo flight, and used large drones for logistics scenarios for the first time. Using drones and unmanned vehicles to deliver goods does not occupy the road resources of the community, greatly alleviates the traffic pressure in the distribution area, and effectively avoids logistics delays caused by traffic jams. Combining drones with express lockers can also reduce working hour restrictions and achieve 24-hour delivery. Customers can arrange their own time to receive packages according to their work and life needs.

[0004] Smart logistics can greatly reduce costs in fuel, manpower, management, carbon emissions, and other aspects, especially for long-distance logistics transportation, where the cost savings are particularly considerable. Smart logistics is an area that my country's logistics industry is actively exploring. Efficient smart logistics has positive significance for improving road transportation capacity, building smart cities, and improving the efficiency of logistics systems and urban resource utilization.

[0005] Traditional intelligent logistics network optimization models and methods have the following disadvantages:

[0006] 1: The computational tasks of the multi-objective optimization problem of the smart logistics network are complex. With the development of smart logistics, the logistics operation objectives have become more complex. In addition to cost being the main objective, carbon emissions are also an objective that must be considered. Compared with developed countries, my country's logistics costs are still too high, and the main reason is that the logistics distribution costs remain high. The traditional model of optimizing internal information construction to improve logistics performance is difficult to achieve satisfactory results. The construction of a smart logistics distribution service system needs to fully consider the sharing of logistics information resources, the sharing of logistics infrastructure equipment, the integration and utilization of logistics resources, and the improvement of the social productivity of the logistics distribution industry, thereby reducing the total logistics cost. In addition, in smart logistics, the site selection and path optimization of distribution centers are still important tasks in the distribution process. How to reduce the site selection cost and path optimization cost in the process of logistics network planning, logistics planning, logistics execution, and logistics recycling is an urgent problem to be solved in the construction of smart distribution. There are many factors to be considered in the design, operation, and optimization of the multi-objective optimization problem of smart logistics, which are divided into two major parts: information system and logistics system. The two are inextricably linked and affect each other. Therefore, the computational tasks of the multi-objective optimization problem of smart logistics are very complex. If the design is not good, it is easy to affect the operational efficiency of smart logistics.

[0007] 2: Traditional methods are difficult to adapt to the high dynamics of smart logistics networks. Traditional logistics optimization methods are fixed conditions for multi-objective optimization problems. In the distribution process, they are based on initial static conditions. They do not consider the dynamic changes of smart logistics networks and are difficult to achieve dynamic calculations. Smart logistics multi-objective optimization problems are dynamic in network planning, logistics planning, logistics execution, and reverse logistics, and often lead to dynamic changes in logistics topology. User nodes and different logistics terminal devices in smart logistics will dynamically join, dynamically exit, and may also fail, resulting in changes in the number of nodes and topology of smart logistics networks. However, the terminal devices used by most smart logistics nodes have limited communication range, and their computing power and storage capacity are insufficient to support complex logistics network optimization methods. Network nodes usually only retain local topology information and cannot obtain dynamic global topology changes at any time. Therefore, traditional logistics network optimization methods are not suitable for multi-objective optimization problems of dynamically changing smart logistics networks, especially large-scale, topology-changing smart logistics network optimization.

[0008] 3: Traditional methods are difficult to adapt to distributed computing in smart logistics networks. Traditional logistics optimization methods often perform offline calculations on a standalone host or server. Complex logistics optimization methods often consume huge amounts of energy. Such methods are not suitable for distributed nodes and smart logistics terminal devices. The energy of smart logistics network nodes is limited, and smart logistics terminal devices are mostly powered by batteries, usually without external energy supplements. Therefore, the multi-objective optimization problem solving method running on the smart logistics network needs to consider energy efficiency and use distributed computing to reasonably utilize the limited computing power and storage capacity of smart logistics terminal devices. Traditional distributed methods need to maintain complex routing tables, and the maintenance and update of routing tables require more computing resources. However, due to the limitations of battery capacity, computing power, and storage capacity, smart logistics nodes are difficult to undertake complex routing table calculation tasks. The network-wide broadcast routing table is also difficult to save due to limited battery capacity and random user exits. Summary of the invention

[0009] The technology proposed in this invention utilizes the limited battery capacity, computing power and storage capacity of smart logistics edge devices. It does not need to maintain complex routing tables and uses probabilistic heuristic methods to improve logistics efficiency while reducing energy consumption and computing resource requirements.

[0010] A multi-objective network optimization method for plant communities in smart logistics edge devices, comprising the following steps:

[0011] Step 1: Plant community solves the smart logistics network planning optimization;

[0012] Step 2: Plant community solves smart logistics plan optimization;

[0013] Step 3, plant community solves smart logistics execution optimization;

[0014] Step 4: Plant community solves the reverse network optimization of smart logistics;

[0015] In step 1, the following sub-steps are included:

[0016] Sub-step 1-1, the logistics manager inputs multiple smart logistics network planning goals on the smart logistics edge device; the smart logistics network planning optimization problem is to solve the optimal set of nodes and edges of the smart logistics network; smart logistics edge devices refer to devices with computing, communication and storage capabilities used in smart logistics operations, including drones, drone remote controls, logistics robots, electric vehicles, mobile computers, portable communication devices, servers, databases, and various types of smart mobile terminals; smart logistics network planning goals include smart logistics network goals, smart logistics transportation tool goals, and smart logistics operation performance goals; smart logistics network goals include the layout of logistics nodes, the connection of logistics edges, the total length of routes or edges, qualitative and quantitative descriptions of the number of logistics nodes, the distance between logistics nodes, storage capacity, and storage costs; smart logistics transportation tool goals include the number of transportation tools, types of transportation tools, transportation tool costs, transportation tool speeds, and transportation tool capacity; smart logistics operation performance goals include service quality levels, carbon emission levels, order completion rates, and customer satisfaction; logistics managers determine computing tasks based on the types and scales of smart logistics and select multiple smart logistics network planning goals;

[0017] In sub-step 1-2, the smart logistics edge device randomly generates a feasible solution for the optimization of the smart logistics network planning; after the logistics manager selects multiple smart logistics network planning targets on the smart logistics edge device, it is transmitted to multiple smart logistics edge devices through distributed communication; each smart logistics edge device that completes the distributed communication randomly generates a feasible solution for the optimization of the smart logistics network planning according to multiple smart logistics network planning targets; further, each smart logistics edge device calculates a feasible solution for the multi-objective optimization of the smart logistics network planning, which corresponds to an individual plant, and the smart logistics edge devices participating in the multi-objective optimization of the smart logistics network planning jointly compare the fitness of the feasible solutions through distributed communication, and jointly select the optimal solution for the multi-objective optimization of the smart logistics network planning with the highest fitness;

[0018] In sub-steps 1-3, the plant community calculates the optimal solution for multi-objective optimization of smart logistics network planning on the smart logistics edge device; the smart logistics edge device sets the multiple smart logistics network planning objectives obtained in sub-step 1-1 as the fitness function, inputs it to the plant community, and sets the feasible solution of smart logistics network planning optimization randomly generated in sub-step 1-2 as the initial plant individual; in the first iterative calculation, the plant community initializes parameters, including setting the size of the plant population in the plant community, the numerical type and numerical size of the plant individual according to the number of edge devices in the smart logistics network planning process, and encoding the plant individual according to the smart logistics network planning optimization, that is, encoding and solving the set of nodes and edges of the smart logistics network; setting the growth variation probability, flowering probability, and fruiting probability of the plant community; initializing the smart logistics network set of plant individuals, clearing the plant sowing set, clearing the plant community flowering set, and clearing the plant community The result pairing set is dropped, and the plant individuals with the highest sorting priority are cleared; further, the starting and ending conditions of the plant community algorithm are initialized, including the calculation start time, the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; in the subsequent iterations, different from the first iteration, the plant community generates a multi-objective feasible solution for the smart logistics network planning on the edge devices in the smart logistics network planning process, calculates the fitness function value of the plant individuals, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individuals with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individuals with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics network planning, that is, the optimal set of smart logistics network nodes and edges;

[0019] In step 2, the following sub-steps are included:

[0020] In sub-step 2-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics network planning; the smart logistics planning optimization problem is to solve the optimal smart logistics planning schedule and transportation tool set; in sub-step 1-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of the smart logistics network planning obtained in sub-step 1-3 to the entire network, and the edge devices receive the broadcast during the smart logistics planning process, and store the optimal solution of the multi-objective optimization of the smart logistics network planning in this node, and clear the multi-objective feasible solutions of the smart logistics network planning that were originally calculated and stored; further, the smart logistics edge device constrains the plant community search according to the optimal solution of the multi-objective optimization of the smart logistics network planning, that is, the plant community searches for a feasible solution of the smart logistics plan optimization in the set of nodes and edges of the optimal smart logistics network; each smart logistics edge device calculates a feasible solution of the multi-objective optimization of the smart logistics plan, that is, corresponding to a plant individual, and the smart logistics edge devices participating in the multi-objective optimization of the smart logistics plan jointly prepare to solve the optimization of the smart logistics plan;

[0021] In sub-step 2-2, the logistics planner inputs multiple smart logistics planning goals on the smart logistics edge device; the logistics planner refers to the optimal solution of multi-objective optimization of smart logistics network planning, inputs the real smart logistics planning goals through the smart logistics edge device, and converts them into smart logistics plans in the virtual world, preferably including multiple smart logistics planning goals; the smart logistics planning goals include the start time and expected deadline of the smart logistics plan, the name of the shipper, the name of the consignee, the shipping address, the receiving address, the code and name of the goods, the type of goods, the quantity of goods, the weight of goods, the distribution network node set, the total number of customers, the set of road sections between nodes, the total number of transport tools, the type of transport tools, the weight limit of transport tools, the transportation mileage, customer customization requirements and additional requests, the time of arrival at the transit logistics node, the time of leaving the transit logistics node, the earliest time for the customer to receive the service, the latest time for the customer to receive the service, the unit time penalty cost incurred by the transport tool for arriving at the customer point in advance, and the unit time penalty cost incurred by the transport tool for arriving at the customer point late time penalty cost, cost per unit distance traveled by the transport, fixed use cost of the transport, real-time load of the transport between customer points, travel speed of the transport between customer points, fuel consumption of the transport, fuel consumption factor, deadweight of the transport, fuel consumption rate of the transport when empty, fuel consumption rate of the transport when fully loaded; the transport refers to various types of logistics vehicles used in smart logistics, including drones, unmanned vehicles, robots, fuel vehicles, electric vehicles, trains, airplanes, and ships; further, the smart logistics plan goal can measure carbon emissions from the macro perspective of a country, region, or industry, conduct statistical analysis on different energy consumption of a country, region, or industry, and infer carbon dioxide emissions based on fuel consumption in statistical data for a certain period of time. Further, multiply the consumption of different fuels in the statistical period by their corresponding fuel consumption factors. Further, further statistics, classification, or summary of different transport vehicles and fuel types are performed to obtain the total amount of carbon dioxide emissions in the statistical period.

[0022] In sub-step 2-3, the plant community calculates the optimal solution of the multi-objective optimization of the smart logistics plan on the smart logistics edge device; the smart logistics edge device uses the multi-objective optimization optimal solution of the smart logistics network planning obtained in sub-step 2-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics plan goals obtained in sub-step 2-2 are used as the fitness function of the plant community, and are input to the plant community; further, in the first iterative calculation, the plant community initializes parameters, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics planning process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics plan optimization, that is, encoding and solving the smart logistics plan progress and the set of transportation tools; each individual plant corresponds to a feasible solution for the multi-objective optimization of the smart logistics plan, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; initializes the smart logistics plan set of the individual plant, clears the plant sowing Set, clear the plant community flowering set, clear the plant community result pairing set, and clear the plant individual with the highest sorting priority; further, initialize the starting and ending conditions of the plant community algorithm, including calculation start time, calculation end time or iteration calculation number limit, and end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates a multi-objective feasible solution for the smart logistics plan on the edge devices of the smart logistics planning process, calculates the fitness function value of the plant individual, and different smart logistics edge devices compare the fitness function value with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics plan, that is, the optimal smart logistics plan schedule and transportation tool set;

[0023] In step 3, the following sub-steps are included:

[0024] In sub-step 3-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics plan; the smart logistics execution optimization problem is to search for the optimal smart logistics execution time period and transportation tool set; in sub-step 2-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of the smart logistics plan obtained in sub-step 2-3 to the entire network, and the edge devices receive the broadcast during the execution of smart logistics, and store the optimal solution of the multi-objective optimization of the smart logistics plan in this node, and clear the feasible solutions of the multi-objective optimization of the smart logistics plan that were originally calculated and stored; further, the smart logistics edge device constrains the plant community search according to the optimal solution of the multi-objective optimization of the smart logistics plan, that is, the plant community searches for a feasible solution of the smart logistics execution optimization in the optimal smart logistics plan schedule and transportation tool set; each smart logistics edge device calculates a feasible solution for the multi-objective optimization of smart logistics execution, that is, corresponding to a plant individual, and the smart logistics edge devices participating in the multi-objective optimization of smart logistics execution jointly prepare to solve the smart logistics execution optimization;

[0025] In sub-step 3-2, the logistics executor inputs multiple smart logistics execution targets on the smart logistics edge device; the logistics executor refers to the optimal solution of the multi-objective optimization of the smart logistics plan, inputs the actual smart logistics execution targets through the smart logistics edge device, and converts them into real-world smart logistics execution, preferably including multiple smart logistics execution targets; the smart logistics execution targets include the name of the shipper, the name of the consignee, the shipping address, the receiving address, the code and name of the goods, the type of goods, the quantity of goods, the weight of the goods, the logistics nodes passed during the transportation process, and the different transportation sections according to the different logistics nodes passed, each transportation section uses different means of transportation, the name and code of the logistics executor responsible for each transportation section, the start time and end time of each transportation section, the signature of each transportation section, and the intact status of the goods in each transportation section;

[0026] In sub-step 3-3, the plant community calculates the optimal solution for multi-objective optimization of smart logistics execution on the smart logistics edge device; the smart logistics edge device uses the optimal solution for multi-objective optimization of the smart logistics plan obtained in sub-step 3-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics execution goals obtained in sub-step 3-2 as the fitness function of the plant community, and inputs them to the plant community; further, in the first iterative calculation, the plant community initializes parameters, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics execution process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics execution optimization, that is, encoding into a smart logistics execution time period and a set of transportation tools; each individual plant corresponds to a feasible solution for multi-objective optimization of smart logistics execution, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; initializes the smart logistics execution set of the individual plant, clears the plant sowing set, Clear the flowering set of the plant community, clear the pairing set of the plant community results, and clear the plant individuals with the highest sorting priority; further, initialize the starting and ending conditions of the plant community algorithm, including the calculation start time, the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates a feasible solution for the multi-objective execution of smart logistics on the edge devices of the smart logistics execution process, calculates the fitness function value of the individual plant, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of smart logistics execution, that is, the optimal smart logistics execution time period and transportation tool set;

[0027] In step 4, the following steps are included:

[0028] In sub-step 4-1, the smart logistics edge device inputs the optimal solution for multi-objective optimization of smart logistics execution; the smart logistics reverse network optimization problem is to solve the optimal set of reverse goods and reverse transportation routes; in sub-step 3-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution for multi-objective optimization of smart logistics execution obtained in sub-step 3-3 to the entire network, and all edge devices in the smart logistics reverse network receive the broadcast, and store the optimal solution for multi-objective optimization of smart logistics execution in this node, and clear the feasible solutions for multi-objective optimization of smart logistics execution that were originally calculated and stored; further, the smart logistics edge device constrains the plant community search according to the optimal solution for multi-objective optimization of smart logistics execution, that is, the plant community searches for a feasible solution for smart logistics reverse optimization in the optimal smart logistics execution time period and transportation tool set; each smart logistics edge device calculates a feasible solution for multi-objective optimization of smart logistics reverse network, that is, corresponding to a plant individual, and the smart logistics edge devices participating in the multi-objective optimization of smart logistics reverse network jointly prepare to solve the smart logistics reverse network optimization;

[0029] In sub-step 4-2, the logistics reverse personnel input multiple smart logistics reverse network optimization goals on the smart logistics edge device; the logistics reverse personnel refer to the smart logistics execution multi-objective optimization optimal solution, input the actual smart logistics reverse goals through the smart logistics edge device, and convert them into real-world smart logistics reverse operations, preferably, including multiple smart logistics reverse goals; the smart logistics reverse goals include the reason for return or exchange, the name of the returner, the name of the consignee, the returner address, the consignment address, the code and name of the returned goods, the intact state of the returned goods, the type of returned goods, the quantity of returned goods, the weight of the returned goods, the logistics nodes passed through the reverse transportation process, and the reverse transportation sections are divided into different reverse transportation sections according to the different reverse logistics nodes passed through, each reverse transportation section uses a different reverse transportation tool, the name and code of the logistics reverse personnel responsible for each reverse transportation section, the start time and end time of each reverse transportation section, the signature of each reverse transportation section, and the intact state of the goods in each reverse transportation section;

[0030] In sub-step 4-3, the plant community calculates the optimal solution of the multi-objective optimization of the smart logistics reverse network on the smart logistics edge device; the smart logistics edge device uses the optimal solution of the multi-objective optimization of smart logistics obtained in sub-step 4-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics reverse goals obtained in sub-step 4-2 are used as the fitness function of the plant community, and are input to the plant community; further, in the first iterative calculation, the plant community initializes the parameters, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics reverse process, the numerical type and numerical size of the individual plant, and the individual plant is encoded according to the smart logistics reverse optimization, that is, encoded into a set of reverse goods and reverse transportation routes; each individual plant corresponds to a feasible solution of the multi-objective optimization of the smart logistics reverse network, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; initializes the smart logistics reverse set of the individual plant, clears the plant sowing set, Clear the flowering set of the plant community, clear the pairing set of the plant community results, and clear the plant individuals with the highest sorting priority; further, initialize the starting and ending conditions of the plant community algorithm, including the calculation start time, the calculation end time or the limit of the number of iterations, and the end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates a multi-objective feasible solution of the smart logistics reverse network on the edge devices of the smart logistics reverse process, calculates the fitness function value of the plant individuals, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individuals with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individuals with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics reverse network, that is, the optimal set of reverse goods and reverse transportation routes.

[0031] Compared with the prior art, the present invention has the following technical effects:

[0032] 1. The present invention is more suitable for solving multi-objective optimization problems of smart logistics networks. This method simultaneously takes into account the characteristics of smart logistics in different stages of network planning, logistics planning, logistics execution, and reverse logistics and the changes in multi-objective optimization problems. This method takes into account the changes in multi-objective optimization problems at different stages, and also takes into account the actual changes and adjustments of logistics managers, logistics planners, logistics executors, and logistics reverse personnel when facing different smart logistics scenarios. The multi-objective optimization problem of the smart logistics network of this method is not static. It allows logistics managers, logistics planners, logistics executors, and logistics reverse personnel to select different smart logistics network optimization goals at different stages of network planning, logistics planning, logistics execution, and reverse logistics, and design different multi-objective optimization problems of smart logistics networks. It has strong flexibility and practical applicability.

[0033] 2. This method can adapt to the high dynamics of smart logistics network optimization. This method mainly completes the calculation on the smart logistics edge device, that is, the device with computing, communication and storage capabilities used in the operation of smart logistics, including drones, drone remote controls, logistics robots, electric vehicles, mobile computers, portable communication devices, servers, databases, and various types of smart mobile terminals. The present invention takes into account the limited resources of smart logistics edge devices and does not use complex loop nesting and a large amount of data exchange and data processing. In the process of smart logistics network planning optimization, smart logistics plan optimization, smart logistics execution optimization, and smart logistics reverse network optimization, the smart logistics edge devices participating in the optimization participate in the calculation each time, that is, the smart logistics edge devices participating in the optimization each time are dynamically changed. This method runs the sowing, growth, flowering, and fruiting operations of the plant community on the smart logistics edge devices participating in the optimization. Different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication, and jointly select the plant individuals with the best fitness, which greatly reduces the time complexity and space complexity of the algorithm.

[0034] 3. This method can realize the distributed computing of multi-objective optimization problems of smart logistics networks, and the method is highly efficient. This method uses the probability search of multiple plant individuals in the plant community to calculate a feasible solution for multi-objective optimization of smart logistics networks on each smart logistics edge device, that is, corresponding to a plant individual. The smart logistics edge devices participating in the multi-objective optimization of the smart logistics network compare the fitness of the feasible solutions through distributed communication, and jointly select the optimal solution for multi-objective optimization of the smart logistics network with the highest fitness, which can be deployed on different smart logistics edge devices for distributed computing. In the design, operation, and optimization process, this method does not need to rely on the establishment, maintenance, update, and search of the global routing table, nor does it require all smart logistics edge devices to participate in the calculation. The amount of calculation is low, the energy emission range requirement is low, and the dynamic joining, exiting, and failure of individual smart logistics edge devices do not affect the calculation results of the entire smart logistics network. The final optimal solution is distributed searched by using a probabilistic plant community-inspired method on the distributed smart logistics edge devices, and the optimization of the smart logistics network is completed through distributed communication and joint election.

[0035] 4. This method has good scalability. This method allows the selection of different smart logistics edge devices in different stages of network planning, logistics planning, logistics execution, and reverse logistics. It does not require all smart logistics edge devices to participate in all operations. Each operation sets the size of the plant population in the plant community according to the number of edge devices in the smart logistics network optimization process. In the design, operation, and optimization process, this method allows logistics managers, logistics planners, logistics executors, and logistics reverse personnel to design smart logistics multi-objective optimization objectives and fitness functions according to the usage scenarios. It can fully consider the influence of actual factors and uncertain environments, including weather changes, traffic conditions, passenger flow changes, traffic accidents, natural disasters, edge device energy limitations, edge device computing power and storage capacity changes, local topology information, dynamic changes, scalability, complexity, and routing update strategies, thereby further improving the applicability of the method in various engineering applications and dynamic scenarios. The process of the present invention is simple, does not require the maintenance of complex routing tables, and does not require complex routing table creation, maintenance, update, and search operations, which provides good scalability for the present invention and allows the present invention to change with the topology of the smart logistics network. As the volume of smart logistics transactions increases and the number of smart logistics nodes increases, the computing performance of the present invention will not deteriorate sharply. Instead, it will improve the overall performance and computing power, making it more suitable for large-scale, uncertain smart logistics network expansion and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0037] Figure 1 It is a flow chart of the working method of the present invention. DETAILED DESCRIPTION

[0038] like Figure 1 As shown, a multi-objective network optimization method for plant communities in smart logistics edge devices includes the following steps:

[0039] Step 1: Solve the smart logistics network planning optimization problem through the plant community algorithm;

[0040] Step 2: Solve the smart logistics planning optimization problem through the plant community algorithm;

[0041] Step 3: Solve the smart logistics execution optimization problem through the plant community algorithm;

[0042] Step 4: Solve the smart logistics reverse network optimization problem through the plant community algorithm;

[0043] The above steps are used to optimize the multi-objective network of plant communities in smart logistics edge devices.

[0044] In step 1, the following sub-steps are included:

[0045] Sub-step 1-1: The logistics manager inputs multiple smart logistics network planning goals on the smart logistics edge device;

[0046] Sub-step 1-2, the smart logistics edge device randomly generates a feasible solution for smart logistics network planning optimization;

[0047] In sub-steps 1-3, the plant community calculates the optimal solution for the multi-objective optimization of the smart logistics network planning on the smart logistics edge device; the smart logistics edge device sets the multiple smart logistics network planning objectives obtained in sub-step 1-1 as the fitness function, inputs it to the plant community, and sets the feasible solution for the smart logistics network planning optimization randomly generated in sub-step 1-2 as the initial plant individual.

[0048] In sub-step 1-1, the smart logistics network planning optimization problem is to solve the optimal set of nodes and edges of the smart logistics network; smart logistics edge devices refer to devices with computing, communication and storage capabilities used in smart logistics operations, including drones, drone remote controls, logistics robots, electric vehicles, mobile computers, portable communication devices, servers, databases, and various types of smart mobile terminals; smart logistics network planning goals include smart logistics network goals, smart logistics transportation tool goals, and smart logistics operation performance goals; smart logistics network goals include the layout of logistics nodes, the connection of logistics edges, the total length of routes or edges, qualitative and quantitative descriptions of the number of logistics nodes, the distance between logistics nodes, storage capacity, and storage costs; smart logistics transportation tool goals include the number of transportation tools, types of transportation tools, transportation tool costs, transportation tool speeds, and transportation tool capacity; smart logistics operation performance goals include service quality levels, carbon emission levels, order completion rates, and customer satisfaction; logistics managers determine computing tasks based on the types and scales of smart logistics and select multiple smart logistics network planning goals;

[0049] In sub-step 1-2, after the logistics manager selects multiple smart logistics network planning targets on the smart logistics edge device, they are transmitted to multiple smart logistics edge devices through distributed communication; each smart logistics edge device that completes the distributed communication randomly generates a feasible solution for smart logistics network planning optimization according to multiple smart logistics network planning targets; further, each smart logistics edge device calculates a feasible solution for multi-objective optimization of smart logistics network planning, which corresponds to an individual plant, and the smart logistics edge devices participating in the multi-objective optimization of smart logistics network planning jointly compare the fitness of feasible solutions through distributed communication, and jointly select the optimal solution for multi-objective optimization of smart logistics network planning with the highest fitness;

[0050] In sub-steps 1-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics network planning process, the numerical type and numerical size of the plant individuals, and encoding the plant individuals according to the smart logistics network planning optimization, that is, encoding and solving the set of nodes and edges of the smart logistics network; setting the growth variation probability, flowering probability, and fruiting probability of the plant community; initializing the smart logistics network set of the plant individuals, clearing the plant sowing set, clearing the plant community flowering set, clearing the plant community result pairing set, and clearing the plant individuals with the highest sorting priority; further, initializing the starting and termination conditions of the plant community algorithm, including when the calculation starts time, calculate the end time or the number of iterations, and the end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates multi-objective feasible solutions for smart logistics network planning on the edge devices in the smart logistics network planning process, calculates the fitness function values ​​of individual plants, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for multi-objective optimization of smart logistics network planning, that is, the optimal set of smart logistics network nodes and edges.

[0051] In step 2, the following sub-steps are included:

[0052] Sub-step 2-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics network planning; specifically: the optimal solution of the multi-objective optimization of the smart logistics network planning obtained in sub-steps 1-3 is broadcast to the entire network, and the edge devices receive the broadcast during the smart logistics planning process, and the optimal solution of the multi-objective optimization of the smart logistics network planning is stored in this node, and the multi-objective feasible solutions of the smart logistics network planning originally calculated and stored are cleared; the smart logistics edge device constrains the plant community search according to the optimal solution of the multi-objective optimization of the smart logistics network planning, that is, the plant community searches for a feasible solution for the optimization of the smart logistics plan in the set of nodes and edges of the optimal smart logistics network; each smart logistics edge device calculates a feasible solution for the multi-objective optimization of the smart logistics plan, which corresponds to an individual plant, and the smart logistics edge devices participating in the multi-objective optimization of the smart logistics plan jointly prepare to solve the optimization of the smart logistics plan;

[0053] In sub-step 2-2, the logistics planner inputs multiple smart logistics planning targets on the smart logistics edge device. Specifically, the logistics planner refers to the optimal solution of multi-objective optimization of smart logistics network planning, inputs the real smart logistics planning targets through the smart logistics edge device, and converts them into smart logistics plans in the virtual world.

[0054] In sub-step 2-3, the plant community calculates the optimal solution of the multi-objective optimization of the smart logistics plan on the smart logistics edge device; specifically: the smart logistics edge device uses the multi-objective optimization optimal solution of the smart logistics network planning obtained in sub-step 2-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics plan goals obtained in sub-step 2-2 are used as the fitness function of the plant community, and input into the plant community.

[0055] In sub-step 2-1, the smart logistics plan optimization problem is to solve the optimal smart logistics plan schedule and transportation tool set; the smart logistics edge device where the plant individual with the highest fitness is located;

[0056] In sub-step 2-2, multiple smart logistics planning goals are included; the smart logistics planning goals include the start time and expected deadline of the smart logistics plan, the name of the shipper, the name of the consignee, the shipping address, the receiving address, the code and name of the goods, the type of goods, the quantity of goods, the weight of goods, the distribution network node set, the total number of customers, the set of road sections between nodes, the total number of transport tools, the type of transport tools, the weight limit of transport tools, the transportation mileage, customer customization requirements and additional requests, the time of arrival at the transit logistics node, the time of departure from the transit logistics node, the earliest time for the customer to receive the service, the latest time for the customer to receive the service, the unit time penalty cost of the transport tool for arriving at the customer point in advance, the unit time penalty cost of the transport tool for arriving at the customer point late, the cost of the transport tool traveling a unit distance, the fixed usage cost of the transport tool, and the cost of the transport tool at Real-time load between customer points, speed of transport vehicles between customer points, fuel consumption of transport vehicles, fuel consumption factor, deadweight of transport vehicles, fuel consumption rate when transport vehicles are empty, fuel consumption rate when transport vehicles are fully loaded; transport vehicles refer to various types of logistics vehicles used in smart logistics, including drones, unmanned vehicles, robots, fuel vehicles, electric vehicles, trains, airplanes, and ships; smart logistics planning goals can measure carbon emissions from a macro perspective of a country, region, or industry, conduct statistical analysis on different energy consumption in a country, region, or industry, and infer carbon dioxide emissions based on fuel consumption in statistical data for a certain period of time; multiply the consumption of different fuels in the statistical period by their corresponding fuel consumption factors; further count, classify or summarize different transport vehicles and fuel types to obtain the total amount of carbon dioxide emissions in the statistical period;

[0057] In sub-steps 2-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics planning process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics plan optimization, that is, encoding and solving the smart logistics plan schedule and the set of transportation tools; each individual plant corresponds to a feasible solution for the multi-objective optimization of the smart logistics plan, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; initialize the smart logistics plan set of the individual plant, clear the plant sowing set, clear the plant community flowering set, clear the plant community result pairing set, and clear the plant individual with the highest sorting priority; further, initialize the starting point of the plant community algorithm Conditions and termination conditions, including calculation start time, calculation end time or iteration calculation number limit, and end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates multi-objective feasible solutions for the smart logistics plan on the edge devices of the smart logistics planning process, calculates the fitness function value of the individual plant, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics plan, that is, the optimal smart logistics plan schedule and transportation tool set.

[0058] In step 3, the following sub-steps are included:

[0059] In sub-step 3-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics plan; specifically: the smart logistics execution optimization problem is to search for the optimal smart logistics execution time period and transportation tool set; in sub-step 2-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of the smart logistics plan obtained in sub-step 2-3 to the entire network. During the execution of smart logistics, all edge devices receive the broadcast and store the optimal solution of the multi-objective optimization of the smart logistics plan in this node, and at the same time clear the feasible solutions of the multi-objective optimization of the smart logistics plan that were originally calculated and stored.

[0060] In sub-step 3-2, the logistics executor inputs multiple smart logistics execution targets on the smart logistics edge device. Specifically, the logistics executor refers to the optimal solution of the multi-objective optimization of the smart logistics plan, inputs the actual smart logistics execution targets through the smart logistics edge device, and converts them into smart logistics execution in the real world.

[0061] In sub-step 3-3, the plant community calculates the optimal solution for multi-objective optimization of smart logistics execution on the smart logistics edge device; specifically: the smart logistics edge device uses the multi-objective optimization optimal solution of the smart logistics plan obtained in sub-step 3-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics execution goals obtained in sub-step 3-2 as the fitness function of the plant community, and inputs them into the plant community.

[0062] In sub-step 3-1, the smart logistics edge device constrains the plant community search according to the optimal solution of the smart logistics plan multi-objective optimization, that is, the plant community searches for a feasible solution for the smart logistics execution optimization in the optimal smart logistics plan schedule and transportation tool set; each smart logistics edge device calculates a feasible solution for the smart logistics execution multi-objective optimization, which corresponds to an individual plant, and the smart logistics edge devices participating in the smart logistics execution multi-objective optimization jointly prepare to solve the smart logistics execution optimization;

[0063] In sub-step 3-2, multiple smart logistics execution targets are included; the smart logistics execution targets include the name of the consignor, the name of the consignee, the shipping address, the receiving address, the code and name of the goods, the type of goods, the quantity of goods, the weight of the goods, the logistics nodes passed during the transportation process, the different transportation sections divided according to the different logistics nodes passed, each transportation section uses different transportation tools, the name and code of the logistics executor responsible for each transportation section, the start time and end time of each transportation section, the signature of each transportation section, and the intact status of the goods in each transportation section;

[0064] In sub-step 3-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics execution process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics execution optimization, that is, encoding into a smart logistics execution time period and a set of transportation tools; each individual plant corresponds to a feasible solution for the smart logistics execution multi-objective optimization, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; initialize the smart logistics execution set of the individual plant, clear the plant sowing set, clear the plant community flowering set, clear the plant community result pairing set, and clear the plant individual with the highest sorting priority; further, initialize the starting conditions of the plant community algorithm and termination conditions, including calculation start time, calculation end time or iteration calculation number limit, and termination error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates smart logistics execution multi-objective feasible solutions on the edge devices in the smart logistics execution process, calculates the fitness function values ​​of individual plants, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of smart logistics execution, that is, the optimal smart logistics execution time period and transportation tool set.

[0065] In step 4, the following steps are included:

[0066] Sub-step 4-1, the smart logistics edge device inputs the optimal solution for the smart logistics multi-objective optimization; the smart logistics reverse network optimization problem, that is, solving the optimal set of reverse goods and reverse transportation routes; in sub-step 3-3, the smart logistics edge device where the plant individual with the highest fitness is located, broadcasts the optimal solution for the smart logistics multi-objective optimization obtained in sub-step 3-3 to the entire network, and all edge devices in the smart logistics reverse network receive the broadcast, and store the optimal solution for the smart logistics multi-objective optimization in this node, and clear the feasible solutions for the smart logistics multi-objective optimization that were originally calculated and stored;

[0067] In sub-step 4-2, the logistics reverse personnel input multiple smart logistics reverse network optimization goals on the smart logistics edge device; specifically: the logistics reverse personnel refer to the smart logistics execution multi-objective optimization optimal solution, input the actual smart logistics reverse goals through the smart logistics edge device, and convert them into the real-world smart logistics reverse operation;

[0068] In sub-step 4-3, the plant community calculates the optimal solution for multi-objective optimization of the smart logistics reverse network on the smart logistics edge device; specifically: the smart logistics edge device uses the optimal solution for multi-objective optimization of smart logistics obtained in sub-step 4-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics reverse objectives obtained in sub-step 4-2 as the fitness function of the plant community, and inputs them into the plant community.

[0069] In step 4-1, the smart logistics edge device constrains the plant community search according to the optimal solution of the smart logistics execution multi-objective optimization, that is, the plant community searches for a feasible solution of the smart logistics reverse optimization in the optimal smart logistics execution time period and transportation tool set; each smart logistics edge device calculates a feasible solution of the smart logistics reverse network multi-objective optimization, that is, corresponding to an individual plant, and the smart logistics edge devices participating in the smart logistics reverse network multi-objective optimization jointly prepare to solve the smart logistics reverse network optimization;

[0070] In step 4-2, multiple smart logistics reverse targets are included; the smart logistics reverse targets include the reason for return or exchange, the name of the returner, the name of the consignee, the address of the returner, the consignee address, the code and name of the returned goods, the intact state of the returned goods, the type of returned goods, the quantity of returned goods, the weight of the returned goods, the logistics nodes passed through during the reverse transportation process, the reverse transportation sections divided according to the different reverse logistics nodes passed through, each reverse transportation section using different reverse transportation tools, the name and code of the logistics reverse personnel responsible for each reverse transportation section, the start time and end time of each reverse transportation section, the signature of each reverse transportation section, and the intact state of the goods in each reverse transportation section;

[0071] In step 4-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics reverse process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics reverse optimization, that is, encoding into a set of reverse goods and reverse transportation routes; each individual plant corresponds to a feasible solution for the multi-objective optimization of the smart logistics reverse network, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; initialize the smart logistics reverse set of individual plant, clear the plant sowing set, clear the plant community flowering set, clear the plant community result pairing set, and clear the plant individual with the highest sorting priority; further, initialize the starting conditions and terminal conditions of the plant community algorithm The stopping conditions include the calculation start time, the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; in the subsequent iterations, different from the first iteration, the plant community generates the multi-objective feasible solution of the smart logistics reverse network on the edge devices of the smart logistics reverse process, calculates the fitness function value of the individual plant, and the different smart logistics edge devices compare the fitness function value with each other through distributed communication. Through continuous sowing, growth, flowering and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics reverse network, that is, the optimal set of reverse goods and reverse transportation routes.

[0072] Example:

[0073] like Figure 1 As shown in the figure, the multi-objective network optimization method of plant communities in smart logistics edge devices completes the multi-objective optimization of smart logistics networks by simulating the seeding, growth, flowering and fruiting behaviors of plant communities on smart logistics edge devices. This method can efficiently complete the multi-objective optimization calculations of smart logistics networks on edge devices.

[0074] The smart logistics network includes smart logistics nodes and smart logistics edges;

[0075] The smart logistics nodes include logistics center nodes that provide materials and customer nodes that need materials; further, logistics center nodes are usually logistics storage centers, distribution centers, transit centers, wholesale centers or retail centers in the logistics network; customer nodes are usually customers located in different geographical locations;

[0076] The smart logistics edge refers to the connection route between smart logistics nodes;

[0077] The smart logistics edge devices refer to devices with computing, communication and storage capabilities used in smart logistics operations, including drones, drone remote controls, logistics robots, electric vehicles, mobile computers, portable communication devices, servers, databases, and various types of smart mobile terminals;

[0078] The plant community is composed of multiple plant individuals, which is a software module running on the smart logistics edge device; further, each smart logistics edge device calculates a feasible solution for the multi-objective optimization of the smart logistics network, that is, corresponding to a plant individual, and the smart logistics edge devices participating in the multi-objective optimization of the smart logistics network jointly compare the fitness of the feasible solution through distributed communication, and jointly select the optimal solution for the multi-objective optimization of the smart logistics network with the highest fitness; further, the plant community includes four operations: sowing, growth, flowering, and fruiting; further, the plant community searches for the optimal solution to the multi-objective optimization problem of the smart logistics network through continuous cycles and iterations of the four operations. Plant sowing set, that is, the set of plant individuals selected for sowing operation; plant community flowering set, that is, the set of plant individuals selected for flowering operation; plant community fruit pairing set, that is, the set of multiple plant individual pairs selected for fruiting operation; the plant with the highest sorting priority indicates that the corresponding plant individual has the highest value of the multi-objective optimization function of the smart logistics network. The growth variation probability of plant communities refers to the probability of plants imitating real plants in nature to have numerical mutations during the growth operation process; the flowering probability of plant communities refers to the probability of plants imitating real plants in nature to select better individuals for flowering operation during the flowering operation process; the fruiting probability of plant communities refers to the probability of plants imitating real plants in nature to carry out cross-learning and fruiting operation during the fruiting process;

[0079] The plant individual refers to a feasible solution to the multi-objective optimization problem of the smart logistics network; further, a plant individual runs on a smart logistics edge device, and its number, i.e., the population size, depends on the number of smart logistics edge devices participating in the optimization;

[0080] The plant community population size, i.e., the number of plant individuals in the plant community, is used for the number of feasible solutions to the multi-objective optimization problem of the smart logistics network; preferably, the number of feasible solutions depends on the number of smart logistics edge devices participating in the optimization, and may also be inconsistent, for example, a smart logistics edge device with strong computing power and storage capacity can also calculate more than two feasible solutions;

[0081] The fitness function of the plant is used to simulate the performance evaluation of solving the multi-objective optimization problem of the smart logistics network. Users can choose different objectives as the multi-objective optimization function of the smart logistics network according to usage needs and different application scenarios, including two aspects of performance evaluation of the information system and the logistics system; the performance of the information system refers to the operating efficiency, operating cost, and energy consumption of the information system of smart logistics in computing, storage, and communication; the performance of the logistics system includes the total distance of the logistics network, the capacity of logistics facilities, transportation tools, transportation distance, transportation volume, transportation speed, transportation cost, transportation price, transportation route, and carbon emissions; further, carbon emissions refer to the calculation and evaluation of greenhouse gas emissions generated during the production, transportation, use and recycling of a product, including the cumulative greenhouse gas emissions per unit of goods. Furthermore, batches of the same product will have different dynamic carbon emissions.

[0082] The sowing operation of the plant community simulates the seed sowing process of a real plant community in nature, and is used to generate the initial feasible solution value required for each round of iterative calculation. The feasible solution value can be used to complete the subsequent growth operation, flowering operation, and fruiting operation; the sowing operation is both random and limited to the vicinity of a certain plant;

[0083] The growth operation of the plant community simulates the growth and mutation process of real plant communities in nature. The sowing operation of plants is random, but only seeds near water sources can continue to grow, and plant individuals that cannot obtain water sources can only die; during the growth process of plants, they may also be affected by water sources or the environment and mutate to become new plant individuals; the growth operation is used to generate new feasible solutions, retaining the main information of the original feasible solutions, but also expanding the search range; the growth mutation probability of the plant community simulates the probability of mutation of real plants in nature during the growth operation, that is, the feasible solution changes in a small range with the mutation probability during the growth process; further, a larger growth mutation probability facilitates the plant community to expand the search space, but it will also lead to a slower convergence speed; a smaller mutation probability is easy to increase the convergence speed of the plant community, but it is also easy to cause the plant community to converge to the local optimal solution earlier;

[0084] The flowering operation of the plant community is used to simulate the flowering process of individual plants in real plant communities in nature. Only branches of individual plants that obtain water and grow vigorously can bloom; the flowering operation is used to retain the better solutions among feasible solutions, that is, the better solutions are retained with a certain flowering probability; the flowering probability of the plant community refers to the probability that the value of the plant plant is selected for the flowering operation during the flowering operation. Not all branches can bloom, and only feasible solutions after flowering can be used for the next result operation; further, a smaller flowering probability facilitates the plant community to expand the search space and retain fewer better solutions, but it will also lead to a slower convergence speed; a larger flowering probability is easy to increase the convergence speed of the plant community and is easy to retain more better solutions of the plant community, but it is also easy to cause the plant community to converge to the local optimal solution earlier;

[0085] The result operation of the plant community is used to simulate the learning function of individual plants in a real plant community in nature, that is, the process in which different plant individuals exchange genetic material through pollen and perform result operation on the optimal solution; the result operation can realize the exchange of data of feasible solutions of different plant individuals, generate new feasible solutions, and simulate the social learning function of the plant community; the result probability of the plant community refers to the probability that different plant individuals learn from each other and perform result operation; a smaller result probability facilitates the plant community to expand the search space, so that the optimal solution retains less, but it will also lead to a slower convergence speed; a larger result probability is easy to increase the convergence speed of the plant community, and it is easy to retain more optimal solutions of the plant community, but it is also easy to cause the plant community to converge to the local optimal solution earlier;

[0086] Furthermore, the multi-objective optimization problem of smart logistics networks refers to the process of rationally optimizing all logistics routes and operation scheduling links to minimize carbon emissions in logistics operations;

[0087] Furthermore, the multi-objective optimization function of the smart logistics network is used to evaluate the quality or performance of feasible solutions to the multi-objective optimization problem of the smart logistics network; further, different logistics processes require different multi-objective optimization functions of the smart logistics network to measure the effectiveness of the smart logistics network optimization from different perspectives, including logistics network planning, logistics planning, logistics execution, and logistics recovery process; further, logistics managers, logistics planners, logistics executors, and logistics reverse personnel can choose different multi-objectives and adjust the multi-objective optimization function according to actual needs and scenario changes when facing different smart logistics scenarios.

[0088] A multi-objective network optimization method for plant communities in smart logistics edge devices includes the following steps:

[0089] Step 1: Plant community solves the smart logistics network planning optimization;

[0090] Step 2: Plant community solves smart logistics plan optimization;

[0091] Step 3, plant community solves smart logistics execution optimization;

[0092] Step 4: Plant community solves the reverse network optimization of smart logistics;

[0093] In step 1, the following sub-steps are included:

[0094] In sub-step 1-1, the logistics manager inputs multiple smart logistics network planning goals on the smart logistics edge device;

[0095] First, referring to traditional logistics problems and combining the characteristics of smart logistics, the following assumptions can be made to facilitate model construction and solution:

[0096] (1) The geographical location, demand, and time window of the logistics center nodes and customer nodes of smart logistics are all known.

[0097] (2) Smart logistics transportation tools are sent out from the distribution center and eventually return to the distribution center.

[0098] (3) Each logistics segment can only be delivered by one means of transport, and the service time requirement remains unchanged.

[0099] (4) All transport vehicles are of the same specifications, each transport vehicle can provide service once, and the total customer demand on each route is less than the maximum load of the transport vehicle.

[0100] (5) There is a soft time window for delivery at the customer point: If the smart logistics transportation tool arrives at the customer point before the delivery time, the service must wait until the delivery time, and a certain waiting cost must be paid; if it arrives later than the delivery time, a delay cost must be paid.

[0101] (6) The time it takes for smart logistics vehicles to leave the distribution center is zero.

[0102] (7) The carbon emissions of smart logistics are related to the operation of transportation vehicles.

[0103] Secondly, this step defines the relevant parameters for the multi-objective optimization problem of smart logistics, and further, mainly includes the following parameters:

[0104] V: A set of smart logistics nodes, V = {1, 2...n}, where 1 represents a distribution center.

[0105] V0: the total number of client nodes, V0 = {2...n}.

[0106] E: The set of edges between smart logistics nodes, E = {(i,j)|i≠j,i,j∈V}.

[0107] n: the total number of transport vehicles, the transport vehicle number k∈{1,2,…n}.

[0108] l: Weight limit for transportation vehicles.

[0109] d ij : The length of the edge from customer point i to j, that is, the distance.

[0110] p i : The demand quantity of customer point i.

[0111] t(i): The time it takes for smart logistics services to reach customers.

[0112] e(i): The earliest time when customers receive smart logistics services.

[0113] l(i): The latest time for customers to receive smart logistics services.

[0114] c1: represents the unit time penalty cost incurred by the smart logistics transportation tool for arriving at the customer point ahead of schedule.

[0115] c2: represents the unit time penalty cost caused by the smart logistics transportation tool's delay in arriving at the customer point.

[0116] c: The cost per unit distance traveled by smart logistics transport vehicles.

[0117] f: represents the fixed usage cost of each smart logistics transportation tool.

[0118] l ij :Smart logistics transportation tools carry real-time load from customer point i to customer point j.

[0119] v ij : The speed of smart logistics transportation tools from customer point i to customer point j.

[0120] C fuel : Fuel consumption of smart logistics transportation vehicles.

[0121] B: Energy consumption factor of smart logistics transportation tools, namely fuel consumption factor and electricity consumption factor.

[0122] t: represents the deadweight of the smart logistics transport vehicle.

[0123] τ0: represents the energy consumption rate of the smart logistics transport vehicle when it is idling, that is, the fuel consumption rate and electricity consumption rate when it is idling.

[0124] τ * : It is the energy consumption rate of the smart logistics transport vehicle when fully loaded, that is, the fuel consumption rate and electricity consumption rate when fully loaded.

[0125]

[0126] Thirdly, the main task of smart logistics network planning is to plan and optimize the nodes and edges of smart logistics, that is, the main computing task is the topology optimization of the smart logistics network, which lays a good network foundation for the subsequent smart logistics planning, execution, and reverse optimization; further, the main computing work of the topology optimization of the smart logistics network includes the capacity, layout, transportation tools, transportation distance, transportation volume, transportation speed, transportation cost, transportation price, and transportation route of logistics facilities; further, the network planning of smart logistics can also consider computing tasks, including the influence of factors such as fuel consumption, speed, load, acceleration, road slope and traffic congestion; further, combining the topology optimization of the smart logistics network with carbon emission optimization can further improve the efficiency of smart logistics network planning and optimization.

[0127] Furthermore, referring to relevant data, the carbon emissions γ1 and fuel consumption F oil There is an obvious relationship. Considering the fuel consumption factor as C, the calculation formula of its carbon emissions is shown in formula (1):

[0128] γ1=C*F oil (1)

[0129] Furthermore, the fuel consumption factor C of different fuels can be set to 1.902kg / L for raw coal, 2.862kg / L for coke, 3.021kg / L for crude oil, 2.544kg / L for gasoline, 2.611kg / L for diesel, and 2.167kg / L for natural gas.

[0130] Furthermore, by referring to the differences in domestic and foreign logistics transportation tools, the combustion efficiency of different fuels and the transportation tool parameters provided by the transportation tool manufacturers, various fuel emission calculation models are constructed with different examples to build a more accurate mathematical model for carbon emissions.

[0131] Sub-step 1-2, the smart logistics edge device randomly generates a feasible solution for smart logistics network planning optimization;

[0132] Furthermore, the smart logistics network planning objectives need to be measured from the macro perspective of a country, region or industry, and statistical analysis should be conducted on the different energy consumption and costs of a country, region or industry, and the network planning should be optimized based on the fuel consumption and cost data in the statistical data of a certain period. Furthermore, the consumption of different fuels and their corresponding fuel consumption factors during the statistical period should be analyzed, and different transportation tools and fuel types should be further counted, classified or summarized, and the transportation tools used in the network planning should be planned in advance.

[0133] Furthermore, according to the positive correlation between carbon emissions and fuel consumption, a positive number is set as the fuel consumption factor to calculate the carbon emissions of the transportation tool; further, according to the differences in domestic and foreign logistics transportation tools, the combustion efficiency of different fuels and the transportation tool parameters provided by the transportation tool manufacturer, various fuel emission calculation models are constructed with different examples to build a more accurate mathematical model for smart logistics network planning;

[0134] Furthermore, logistics managers select and optimize network planning objectives according to the actual operation of smart logistics, policy adjustments, and environmental changes; preferably, smart logistics network planning objectives include reasonably determining nodes and edges of smart logistics, the total length of logistics routes or edges, and qualitatively and quantitatively describing the number of logistics nodes, the distance between logistics nodes, storage capacity, storage costs, the number of transportation tools, types of transportation tools, transportation tool costs, transportation tool speeds, transportation tool capacity, and service quality levels; further, smart logistics network planning objectives are not fixed, and logistics management can re-select smart logistics network planning objectives according to computing task requirements in actual logistics operations;

[0135] Encode the smart logistics network planning optimization problem solved by the plant community algorithm; encode individual plants according to logistics nodes and logistics edges, the encoding length depends on the number of logistics nodes and logistics edges, and the encoded value depends on whether the logistics nodes and logistics edges are selected as feasible solutions by the individual plants; preferably, 1 indicates that the logistics nodes and logistics edges are selected as feasible solutions by the smart logistics edge device, and conversely, 0 indicates that the logistics nodes and logistics edges are not selected as feasible solutions by the individual plants; during initialization, a plant seeding set is randomly generated according to the plant community population size, and each plant individual in the seeding set is generated as a random number, and the number of plant individuals in the seeding set is the plant community population size;

[0136] In sub-step 1-3, the plant community calculates the optimal solution of multi-objective optimization of smart logistics network planning on the smart logistics edge device; the smart logistics edge device sets the multiple smart logistics network planning objectives obtained in sub-step 1-1 as the fitness function, inputs it to the plant community, and sets the feasible solution of smart logistics network planning optimization randomly generated in sub-step 1-2 as the initial plant individual;

[0137] Furthermore, the seeds used in the seeding operation of the first iteration and the seeding operation of subsequent iterations are different; in the first iteration calculation, the seeds of the seeding operation are randomly generated feasible solutions for multi-objective optimization of smart logistics network planning; in subsequent iterations, the seeds in the seeding operation are the fruits produced in the result operation of the previous iteration, that is, the optimal feasible solution for multi-objective optimization of smart logistics network planning;

[0138] Furthermore, from a macroscopic perspective, the energy statistics method is mainly used to estimate carbon dioxide emissions based on fuel consumption. The consumption of different fuels in a certain period of time is multiplied by the corresponding fuel consumption factor, and then the total amount of carbon dioxide emissions (CO2) in this period of time is obtained through statistics. The mathematical expression is:

[0139]

[0140] In formula (2), S is the total CO2 emission; j is the fuel type; ρ j is the fuel conversion factor of the jth fuel; N j is the jth fuel consumption. Through this model, the CO2 emissions can be inferred based on the consumption of different types of fuel.

[0141] Furthermore, in the planning of smart logistics networks, by analyzing the factors that affect the fuel consumption of transportation tools, including the speed of transportation tools, the distance traveled by transportation tools, the type of fuel, the engine traction, the acceleration, etc., an instantaneous fuel consumption carbon emission calculation model is constructed. This method can directly calculate the instantaneous fuel consumption of transportation tools under various working conditions and the actual fuel consumption per unit time without considering the quality of the transportation tools themselves. Furthermore, assuming that p(t) represents the fuel consumption per unit time (ml / s); a represents the acceleration of the transportation tool (m / s 2 ); τ t represents the capacity of the transport vehicle during driving (KN), and the instantaneous fuel consumption carbon emission calculation model is shown in formula (3):

[0142]

[0143] Assume that α represents the fuel consumption rate of the vehicle when it is traveling at a constant speed (ml / s); μ1 represents the fuel consumed per unit time (ml / kJ); μ2 represents the fuel consumption per unit time per second (ml / (kJ / s 2 )); t1 is rolling resistance (KN); t2 is air resistance (ml / (K / s 2 )); η is a percentage; T is the total weight of the vehicle (kg); v is the speed of the vehicle (m / s). Then the capacity of the vehicle when traveling (KN)τ t The formula is as follows (4):

[0144] τ t =t1+t2v 2 +Tα / 1000+gTη / 1000 (4)

[0145] Then the fuel consumption of the transport tool in time t can be calculated as:

[0146]

[0147] Since there is a certain gap between the instantaneous emission calculation model and the actual situation, it is usually used for transportation tool factor analysis under ideal conditions and for quick calculation reference.

[0148] Furthermore, in the growth operation, a single plant individual performs a growth mutation operation on one or several selected feasible solutions of the multi-objective optimization of smart logistics network planning during the growth process, and reversely modifies the corresponding binary bit value, that is, changes 0 to 1, or vice versa, changes 1 to 0; all smart logistics nodes and edges are re-encoded into plant individuals as feasible solutions to the multi-objective optimization problem of smart logistics network planning after mutation; further, the smart logistics edge device performs distributed communication and compares the fitness of the feasible solutions of the multi-objective optimization problem of smart logistics network planning of the plant individuals; further, the smart logistics edge device sorts all plant individuals according to the values ​​of the multi-objective optimization function of the smart logistics network planning; further, the smart logistics edge device selects the plant individuals with higher calculated values ​​of the multi-objective optimization fitness function of the logistics network planning for the next step of calculation, otherwise, the plant individuals with lower calculated values ​​of the multi-objective optimization function of the smart logistics network are abandoned, that is, they die in the growth operation;

[0149] Furthermore, in the flowering operation, all smart logistics edge devices perform distributed communication, and the plant community selects the better plant individuals of the smart logistics network planning multi-objective optimization function according to the flowering probability; preferably, a higher sorting priority has a higher flowering probability and is more likely to be selected; conversely, a lower sorting priority has a lower flowering probability and is less likely to be selected; further, the smart logistics edge device adds the selected single plant individuals of higher priority of the smart logistics network planning multi-objective optimization function to the plant community flowering set for the next plant community fruiting operation; conversely, the plant individuals that are not selected into the flowering set are abandoned, that is, they die in the flowering operation;

[0150] Furthermore, in the result operation, all smart logistics edge devices perform distributed communication, and each smart logistics edge device selects a pair of plant individuals with a better smart logistics network planning multi-objective optimization function to form a plant community result pairing, and the two paired plant individuals exchange a part of the binary bits with each other, and modify the combined state of the logistics nodes and edges, and reconnect the logistics network into plant individuals, that is, the feasible solution of the smart logistics network planning multi-objective optimization problem, and calculate the smart logistics network planning multi-objective optimization function of the plant individuals;

[0151] Furthermore, sub-steps 1-3 are continuously looped until all smart logistics edge devices select the optimal solution that meets the iteration number requirement or error requirement, and the multi-objective optimization problem of smart logistics network planning is solved;

[0152] In step 2, the following sub-steps are included:

[0153] In sub-step 2-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics network planning; in sub-step 1-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of the smart logistics network planning obtained in sub-step 1-3 to the entire network. During the smart logistics planning process, all edge devices receive the optimal solution broadcast, and use the logistics nodes and edges corresponding to the optimal solution as the network planning that is prioritized in the smart logistics plan, and give priority to the logistics nodes and edges corresponding to the optimal solution in the smart logistics plan, ensuring that the optimal smart logistics plan is arranged under the optimal smart logistics network planning, thereby improving the effectiveness and work efficiency of the logistics plan;

[0154] In sub-step 2-2, the logistics planner inputs multiple smart logistics planning targets on the smart logistics edge device;

[0155] Furthermore, the goal of smart logistics planning needs to understand the historical data and statistical characteristics of logistics during the smart logistics planning process, reasonably arrange and optimize the geographical location, demand, and time window of distribution centers and customer demand points, reasonably arrange transportation tools and service times, and optimize carbon emissions so that subsequent logistics execution can be carried out efficiently;

[0156] Furthermore, the parameters that can be defined in the smart logistics plan include the set of distribution network nodes, the set of the total number of customers, the set of road sections between nodes, the set of the total number of transport vehicles, the weight limit of transport vehicles, the distance between customer points, the demand of customer points, the time of arrival at the customer, the earliest time for the customer to receive service, the latest time for the customer to receive service, the unit time penalty cost of the logistics transport vehicle due to early arrival at the customer point, the unit time penalty cost of the logistics transport vehicle due to delayed arrival at the customer point, the cost of the transport vehicle traveling a unit distance, the fixed use cost of each vehicle, the real-time load of the transport vehicle between customer points, the travel speed of the transport vehicle between customer points, the fuel consumption of the transport vehicle, the fuel consumption factor, the dead weight of the transport vehicle, the fuel consumption rate of the transport vehicle when it is empty, and the fuel consumption rate of the transport vehicle when it is fully loaded;

[0157] Encode the smart logistics planning optimization problem solved by the plant community algorithm; encode individual plants according to the planned schedule and transportation tools, the encoding length depends on the planned schedule and the number of transportation tools, and the encoded value depends on whether the planned schedule and the transportation tools are selected as feasible solutions by the individual plants; preferably, 1 indicates that the planned schedule and the transportation tools are selected as feasible solutions by the smart logistics edge device, and conversely, 0 indicates that the planned schedule and the transportation tools are not selected as feasible solutions by the individual plants; during initialization, a plant seeding set is randomly generated according to the plant community population size, and each plant individual in the seeding set is generated as a random number, and the number of plant individuals in the seeding set is the plant community population size;

[0158] In sub-step 2-3, the plant community calculates the optimal solution of the multi-objective optimization of the smart logistics plan on the smart logistics edge device; the smart logistics edge device uses the optimal solution of the multi-objective optimization of the smart logistics network planning obtained in sub-step 2-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics plan goals obtained in sub-step 2-2 as the fitness function of the plant community, which are input to the plant community;

[0159] Furthermore, the seeds used in the seeding operation of the first iteration and the seeding operation of subsequent iterations are different; in the first iteration calculation, the seeds of the seeding operation are randomly generated feasible solutions for multi-objective optimization of smart logistics plans; in subsequent iterations, the seeds in the seeding operation are the fruits produced in the result operation of the previous iteration, that is, the optimal feasible solution for multi-objective optimization of smart logistics plans;

[0160] Furthermore, a second-order polynomial calculation model based on power is constructed in the smart logistics plan; after investigating the power of transportation tools, two second-order polynomial fuel consumption / carbon emission calculation models can be established. There are mainly two models: the function between the driving speed of the transportation tool and the acceleration level forms the power used by the transportation tool in the first model; the second model requires both external data and engine data, so a modeling method can be used to establish a calculation model for low-carbon driving. The modeling is as follows:

[0161]

[0162] In formula (6), Ω0, Ω1, Ω2, and is the specific index of each vehicle type, w(t) is the total power transmitted by the transportation tool at time t, and the calculation model is:

[0163] w(t)=((R(t)+1.04ma(t)) / 3600β d )v(t) (7)

[0164] Among them, R(t) is the total resistance, βd Indicates transmission efficiency. The formula for calculating total resistance is:

[0165] R(t)=(w / 25.9)L d +L h Hv 2 (t)+98066mL r {c1v(t)+c2}+9.8066mG(t) (8)

[0166] Furthermore, the specific coefficient settings can refer to relevant literature.

[0167] Sub-step 2-3c: Establish a comprehensive emission estimation model.

[0168] Furthermore, the comprehensive emission model includes three modules: engine power module, engine speed module and fuel consumption module. The mathematical expression of traction is as follows:

[0169] F trac =(Ma+Mgsinθ+0.5L d ρHv 2 +MgL r cosθ)v / 1000 (9)

[0170] In formula (9), v represents the speed of the vehicle (m / s), M represents the weight of the vehicle (kg), and L d It represents the traction coefficient of the load-carrying vehicle, and H represents the windward surface area of ​​the vehicle (m 2 ), ρ represents the air density (kg / m 3 ). g is the gravitational constant (m / s 2 ), L r is the air resistance, so the relationship between the engine power of the transport vehicle can be obtained:

[0171] p=F tract / ξ trans +p loss (10)

[0172] In the above formula, p is the engine power (kw / s), ξ trans is the efficiency of the vehicle transmission, p loss It is the power lost by the engine during rotation.

[0173] Furthermore, the relationship between the engine speed and the vehicle speed is:

[0174] V speed =C(R(L) / R(L s ))v (11)

[0175] In the above formula, V speedIndicates the engine speed, C indicates the vehicle is in running state L s The ratio between the vehicle speed and the engine speed, R(L), is when L=1,2,......,L s is the gear ratio value at that time, and v is the travel speed of the transport vehicle (m / s).

[0176] Furthermore, the fuel consumption model:

[0177] F fuel =λ(σV speed d s +P / η) / 44 (12)

[0178] In the above formula, λ represents the ratio between the fuel and air mass, σ represents the engine friction factor, and d S is the engine emissions.

[0179] Furthermore, the comprehensive model can comprehensively consider the causes of fuel consumption during the operation of transportation tools from all aspects, and in actual use, it can be adjusted according to the parameters of different transportation tools.

[0180] Furthermore, in the growth operation, a single plant individual performs a growth mutation operation on one or several selected feasible solutions of the multi-objective optimization of the smart logistics plan during the growth process, and reversely modifies the corresponding binary bit value, that is, changes 0 to 1, or vice versa, changes 1 to 0; all smart logistics plan progress and transportation tools are re-encoded into plant individuals as feasible solutions to the multi-objective optimization problem of the smart logistics plan; further, the smart logistics edge device performs distributed communication and compares the fitness of the feasible solutions of the multi-objective optimization problem of the smart logistics plan of the plant individuals; further, the smart logistics edge device sorts all plant individuals according to the values ​​of the multi-objective optimization function of the smart logistics plan; further, the smart logistics edge device selects plant individuals with higher calculated values ​​of the multi-objective optimization fitness function of the logistics plan for the next step of calculation, otherwise, plant individuals with lower calculated values ​​of the multi-objective optimization function of the smart logistics plan are abandoned, that is, they die in the growth operation;

[0181] Furthermore, in the flowering operation, all smart logistics edge devices perform distributed communication, and the plant community selects the better plant individuals of the smart logistics plan multi-objective optimization function according to the flowering probability; preferably, a higher sorting priority has a higher flowering probability and is more likely to be selected; conversely, a lower sorting priority has a lower flowering probability and is less likely to be selected; further, the smart logistics edge device adds the selected single plant individuals of higher priority of the smart logistics plan multi-objective optimization function to the plant community flowering set for the next plant community fruiting operation; conversely, the plant individuals that are not selected into the flowering set are abandoned, that is, they die in the flowering operation;

[0182] Furthermore, in the result operation, all smart logistics edge devices perform distributed communication, and each smart logistics edge device selects a pair of plant individuals with a better smart logistics plan multi-objective optimization function to form a plant community result pairing, and the two paired plant individuals exchange a part of the binary bits with each other, and modify the logistics plan progress and the combined state of the transportation tools, and reconnect the logistics plan into plant individuals, that is, the feasible solution of the smart logistics plan multi-objective optimization problem, and calculate the smart logistics plan multi-objective optimization function of the plant individuals;

[0183] Furthermore, sub-steps 2-3 are continuously looped until all smart logistics edge devices select the optimal solution that meets the iteration number requirement or error requirement, and the multi-objective optimization problem of the smart logistics plan is solved;

[0184] In step 3, the following sub-steps are included:

[0185] Sub-step 3-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics plan; in sub-step 2-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of the smart logistics plan obtained in sub-step 2-3 to the entire network. During the execution of smart logistics, all edge devices receive the optimal solution broadcast, and use the logistics plan schedule and transportation tools corresponding to the optimal solution as the logistics plan that is prioritized in the execution of smart logistics, and give priority to the logistics plan schedule and transportation tools corresponding to the optimal solution in the execution of smart logistics, to ensure that the optimal smart logistics execution is arranged under the optimal smart logistics plan, thereby improving the effectiveness and work efficiency of logistics execution;

[0186] In sub-step 3-2, the logistics executor inputs multiple smart logistics execution targets on the smart logistics edge device;

[0187] Furthermore, the smart logistics execution target can build an instantaneous fuel consumption carbon emission calculation model by analyzing the factors that affect the fuel consumption of transportation tools, and calculate the carbon emissions of transportation tools by analyzing the transportation tool's driving speed, transportation tool's driving distance, fuel type, engine traction, acceleration, etc.; further, collect the transportation tool's engine power data and emission data, analyze the relationship curve between the two sets of data to establish a second-order polynomial fuel consumption / carbon emission calculation model, including the relationship between the transportation tool's driving speed and acceleration level, the relationship between engine fuel consumption or power consumption and driving behavior, and the relationship between engine power and engine speed and fuel or power consumption, to calculate the optimal smart logistics execution target;

[0188] Encode the smart logistics execution optimization problem solved by the plant community algorithm; encode individual plants according to the execution time period and transportation tools, the encoding length depends on the execution time period and the number of transportation tools, and the encoded value depends on whether the execution time period and the transportation tool are selected as feasible solutions by the individual plants; preferably, 1 indicates that the execution time period and the transportation tool are selected as feasible solutions by the smart logistics edge device, and conversely, 0 indicates that the execution time period and the transportation tool are not selected as feasible solutions by the individual plants; during initialization, a plant seeding set is randomly generated according to the plant community population size, and each plant individual in the seeding set is generated as a random number, and the number of plant individuals in the seeding set is the plant community population size;

[0189] Furthermore, the carbon dioxide emitted by vehicles during operation is often considered to be related to fuel consumption, and fuel consumption is affected by many factors such as the vehicle's speed, vehicle acceleration, road traffic conditions, driving behavior, fuel type, and the vehicle's own weight.

[0190] In sub-step 3-3, the plant community calculates the optimal solution for multi-objective optimization of smart logistics execution on the smart logistics edge device; the smart logistics edge device uses the optimal solution for multi-objective optimization of the smart logistics plan obtained in sub-step 3-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics execution goals obtained in sub-step 3-2 as the fitness function of the plant community, and inputs them to the plant community;

[0191] Furthermore, the seeds used in the seeding operation of the first iteration and the seeding operation of subsequent iterations are different; in the first iteration calculation, the seeds of the seeding operation are randomly generated feasible solutions for the multi-objective optimization of smart logistics; in subsequent iterations, the seeds in the seeding operation are the fruits produced in the result operation of the previous iteration, that is, the optimal feasible solution for the multi-objective optimization of smart logistics;

[0192] Furthermore, a multi-objective optimization model for logistics carbon emissions is proposed in the smart logistics execution, and its optimization goal is to consider Carbon emissions minimize;

[0193] Furthermore, the carbon emission objective function is constructed according to the operation status of the transportation tool; considering the common mode of carbon emission, the carbon emission of the transportation tool and fuel consumption C fuel There is a multiple relationship between them, so a fuel consumption factor B can be used to calculate carbon emissions, that is, the amount of carbon dioxide emitted per unit of fuel consumed. The expression is as follows:

[0194]

[0195] Refer to the comprehensive fuel model established in sub-step 2-3c, where the vehicle load constraint is l, the weight of the vehicle is t, and d ij It refers to the distance between the transport tool from customer point i to customer point j. The real-time load is l ij , the speed of the transport is v ij , the road inclination is θ, F r is air resistance, λ1=0.5F d Sρ,F d is the traction coefficient of the load-carrying vehicle, S is the front surface area of ​​the load-carrying vehicle, and ρ is the air density. The calculation expression of fuel consumption is as follows:

[0196]

[0197] In formula (14), α ij =a+gsinθ+gC r cosθ, a is the acceleration of the vehicle, g is the gravitational constant (m / s 2 ).

[0198] Furthermore, the model can comprehensively consider the impact of factors such as vehicle speed and load on vehicle fuel consumption. Furthermore, according to the actual operation conditions on different roads, the parameters can be further adjusted to better reflect the various energy consumption parameters of the vehicle during driving. Furthermore, the model calculation process is relatively complicated, and it can be simplified according to the impact of load on fuel consumption, introducing the fuel consumption rate of the vehicle when empty is l0 and the fuel consumption rate of the vehicle when fully loaded is l. * , when the load of the transport vehicle is l ij The calculation expression of fuel consumption is:

[0199] C(l ij )=α(t+l ij )+λ (15)

[0200] Assume that the fuel consumption of the transport vehicle when empty is τ0=αt+λ, and the fuel consumption when fully loaded is τ * =α(t+l)+λ, then Substituting into formula (15), the formula is rewritten as:

[0201]

[0202] The effect of load on fuel consumption of transportation tools is replaced by equation (16) in equation (14), and the carbon emission expression is:

[0203]

[0204] Furthermore, in the growth operation, a single plant individual performs a growth mutation operation on one or several selected feasible solutions for multi-objective optimization of smart logistics during the growth process, and reversely modifies the corresponding binary bit value, that is, changes 0 to 1, or vice versa, changes 1 to 0; all smart logistics execution time periods and transportation tools are re-encoded into plant individuals as feasible solutions to the mutated smart logistics execution multi-objective optimization problem; further, the smart logistics edge device performs distributed communication and compares the fitness of the feasible solutions of the smart logistics execution multi-objective optimization problem of the plant individuals; further, the smart logistics edge device sorts all plant individuals according to the values ​​of the smart logistics execution multi-objective optimization function; further, the smart logistics edge device selects plant individuals with higher values ​​calculated by the logistics execution multi-objective optimization fitness function for the next step of calculation, otherwise, plant individuals with lower values ​​calculated by the smart logistics execution multi-objective optimization function are abandoned, that is, they die in the growth operation;

[0205] Furthermore, in the flowering operation, all smart logistics edge devices perform distributed communication, and the plant community selects the better plant individuals for the smart logistics execution multi-objective optimization function according to the flowering probability; preferably, a higher sorting priority has a higher flowering probability and is more likely to be selected; conversely, a lower sorting priority has a lower flowering probability and is less likely to be selected; further, the smart logistics edge device adds the selected single plant individuals with a higher priority for the smart logistics execution multi-objective optimization function to the plant community flowering set for the next plant community fruiting operation; conversely, the plant individuals that are not selected into the flowering set are abandoned, that is, they die in the flowering operation;

[0206] Furthermore, in the result operation, all smart logistics edge devices perform distributed communication, and each smart logistics edge device selects a pair of plant individuals with better smart logistics execution multi-objective optimization function to form a plant community result pairing, and the two paired plant individuals exchange a part of the binary bits with each other, and modify the combined state of the logistics execution time period and the transportation tool, and reconnect the logistics execution into plant individuals, that is, the feasible solution of the smart logistics execution multi-objective optimization problem, and calculate the smart logistics execution multi-objective optimization function of the plant individuals;

[0207] Furthermore, sub-step 3-3 is continuously looped until all smart logistics edge devices select the optimal solution that meets the iteration number requirement or error requirement, and the solution of the multi-objective optimization problem of smart logistics execution is completed;

[0208] In step 4, the following steps are included:

[0209] Sub-step 4-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of smart logistics execution; in sub-step 3-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of smart logistics execution obtained in sub-step 3-3 to the entire network, and the edge devices receive the optimal solution broadcast during the smart logistics reverse process, and use the logistics execution time period and transportation tools corresponding to the optimal solution as the logistics plan that is prioritized in the smart logistics reverse, and give priority to the logistics execution time period and transportation tools corresponding to the optimal solution in the smart logistics reverse, to ensure that the optimal smart logistics reverse is arranged under the optimal smart logistics execution, thereby improving the effectiveness and work efficiency of the logistics reverse;

[0210] In sub-step 4-2, the logistics reverse personnel input multiple smart logistics reverse network optimization goals on the smart logistics edge device;

[0211] Furthermore, the multi-objective priority of the smart logistics reverse network uses smart logistics to execute the optimal solution of multi-objective optimization, that is, to make full use of the forward logistics transportation tools or the return transportation tools to complete the reverse logistics transportation, greatly reducing the cost of reverse logistics transportation; further, according to the characteristics of the reverse goods, determine the appropriate transportation tool load constraints, the weight of the transportation tool, the distance between the transportation tool and the customer point, the real-time load, the transportation tool speed, the road inclination, the air resistance, the traction coefficient of the load-carrying transportation tool, the front surface area of ​​the load-carrying transportation tool, the air density, calculate the fuel consumption, the carbon emission, and calculate the multi-objective function according to the minimum total cost of reverse logistics;

[0212] Furthermore, smart logistics reverse can consider multiple objectives in addition to cost; further, in smart logistics reverse, soft time windows and capacity constraints are considered, the driving speed and load of the transport tool are optimized, the functional relationship of the impact of the delivery distance on fuel consumption is calculated, the total delivery performance is calculated, and other logistics objectives are optimized while considering the reverse total cost minimization objective;

[0213] Encode the plant community algorithm to solve the reverse optimization problem of smart logistics; encode individual plants according to reverse goods and reverse transportation routes, the encoding length depends on the number of reverse goods and reverse transportation routes, and the encoding value depends on whether the reverse goods and reverse transportation routes are selected as feasible solutions by the individual plants; preferably, 1 indicates that the reverse goods and reverse transportation routes are selected as feasible solutions by the smart logistics edge device, and conversely, 0 indicates that the reverse goods and reverse transportation routes are not selected as feasible solutions by the individual plants; during initialization, a plant seeding set is randomly generated according to the plant community population size, and each plant individual in the seeding set is generated as a random number, and the number of plant individuals in the seeding set is the plant community population size;

[0214] In sub-step 4-3, the plant community calculates the optimal solution for the multi-objective optimization of the smart logistics reverse network on the smart logistics edge device; the smart logistics edge device uses the optimal solution for the multi-objective optimization of smart logistics obtained in sub-step 4-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics reverse objectives obtained in sub-step 4-2 as the fitness function of the plant community, and inputs them to the plant community; when the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics reverse network.

[0215] Furthermore, the seeds used in the seeding operation of the first iteration and the seeding operation of subsequent iterations are different; in the first iteration calculation, the seeds of the seeding operation are randomly generated feasible solutions for multi-objective optimization of the smart logistics reverse network; in subsequent iterations, the seeds in the seeding operation are the fruits produced in the result operation of the previous iteration, that is, the optimal feasible solution for multi-objective optimization of the smart logistics reverse network;

[0216] Furthermore, in the multi-objective optimization of the smart logistics reverse network, the soft time window and capacity constraints are considered for optimization. Considering the driving speed, load, and delivery distance of the transport tool, they have an impact on fuel consumption. is the total delivery cost minimize.

[0217] Furthermore, the total cost objective function of reverse delivery of transport tools is constructed

[0218] Furthermore, considering the total delivery cost of logistics transportation tools in the reverse delivery process Mainly fixed costs for the use of transportation tools Cost of distance traveled by transport means and penalty costs for violating customer service time windows composition.

[0219] Furthermore, the fixed cost of reverse transport is constructed, which is the cost during the operation of reverse transport, which is not dependent on the distance traveled by the transport, but is related to the driver's usage fee, loss fee, rental cost, etc. Assuming f is the usage cost of each transport, and n is the number of transports, the fixed cost of transport is calculated as follows:

[0220]

[0221] Furthermore, the reverse transportation tool travel distance cost is constructed, which is the distribution cost consumed by the transportation tool during the distribution process. Assuming c represents the distribution cost per unit distance, the calculation formula for the transportation tool travel distance cost is:

[0222]

[0223] Furthermore, we construct the penalty cost for violating the time window. Considering the logistics problem with time windows, we can see that all reverse transport tools need to pay penalty costs when they exceed the customer's time window. The calculation formula is:

[0224]

[0225] Furthermore, the objective function of the total cost of reverse delivery is constructed as:

[0226]

[0227] Furthermore, a multi-objective optimization model of reverse logistics carbon emissions and total reverse distribution costs is constructed, which is a two-objective model.

[0228]

[0229] Furthermore, through the analysis of carbon emissions and total distribution costs, the two objectives can be combined to construct a multi-objective optimization model for carbon emissions in reverse logistics:

[0230]

[0231]

[0232] Formula (23) represents the objective function of minimizing the carbon emissions generated during the reverse transport delivery process; under certain constraints, the logistics reverse delivery path model is established with the minimum total mileage and time as optimization variables; Formula (24) represents the objective function of minimizing the total delivery cost during the reverse transport delivery; Formula (25) represents the load limit of the customer point; Formula (26) means that each customer point is visited only once; where (27) and (28) mean that each demand point has only one vehicle for delivery; Formula (29) means that the starting point of the reverse transport is the distribution center; Formula (30) indicates that the reverse transport leaves the distribution center at time 0; Formula (31) represents the time window constraint; Formula (32) represents the non-negative constraint.

[0233] Furthermore, in the growth operation, a single plant individual performs a growth mutation operation on one or several selected feasible solutions of the multi-objective optimization of the smart logistics reverse network during the growth process, and reversely modifies the corresponding binary bit value, that is, changes 0 to 1, or vice versa, changes 1 to 0; all smart logistics reverse goods and reverse transportation routes are re-encoded into plant individuals as feasible solutions to the multi-objective optimization problem of the smart logistics reverse network; further, the smart logistics edge device performs distributed communication and compares the fitness of the feasible solutions of the multi-objective optimization problem of the smart logistics reverse network of the plant individuals; further, the smart logistics edge device sorts all plant individuals according to the values ​​of the multi-objective optimization function of the smart logistics reverse network; further, the smart logistics edge device selects plant individuals with higher values ​​calculated by the fitness function of the logistics reverse multi-objective optimization for the next step of calculation, otherwise, plant individuals with lower values ​​calculated by the smart logistics multi-objective optimization function are abandoned, that is, they die in the growth operation;

[0234] Furthermore, in the flowering operation, all smart logistics edge devices perform distributed communication, and the plant community selects the better plant individuals of the smart logistics reverse network multi-objective optimization function according to the flowering probability; preferably, a higher sorting priority has a higher flowering probability and is more likely to be selected; conversely, a lower sorting priority has a lower flowering probability and is less likely to be selected; further, the smart logistics edge device adds the selected single plant individuals with higher priority of the smart logistics reverse network multi-objective optimization function to the plant community flowering set for the next step of the plant community fruiting operation; conversely, the plant individuals that are not selected into the flowering set are abandoned, that is, they die in the flowering operation;

[0235] Furthermore, in the result operation, all smart logistics edge devices perform distributed communication, and each smart logistics edge device selects a pair of plant individuals with a better smart logistics reverse network multi-objective optimization function to form a plant community result pairing, and the two paired plant individuals exchange a part of the binary bits with each other, and modify the combination state of the logistics reverse goods and the reverse transportation route, and reconnect the logistics reverse to the plant individual, that is, the smart logistics executes the feasible solution of the multi-objective optimization problem, and calculates the smart logistics reverse network multi-objective optimization function of the plant individual;

[0236] Furthermore, sub-steps 4-3 are continuously looped until all smart logistics edge devices select the optimal solution that meets the iteration number requirement or error requirement, thus completing the solution of the multi-objective optimization problem of the smart logistics reverse network.

Claims

1. A multi-objective network optimization method for plant communities in smart logistics edge devices, characterized in that: It includes the following steps: Step 1: Solve the smart logistics network planning optimization problem through the plant community algorithm; Step 2: Solve the smart logistics planning optimization problem through the plant community algorithm; Step 3: Solve the smart logistics execution optimization problem through the plant community algorithm; Step 4: Solve the smart logistics reverse network optimization problem through the plant community algorithm; The above steps are used to optimize the multi-objective network of plant communities in smart logistics edge devices.

2. The method according to claim 1, characterized in that In step 1, the following sub-steps are included: Sub-step 1-1: The logistics manager inputs multiple smart logistics network planning goals on the smart logistics edge device; Sub-step 1-2, the smart logistics edge device randomly generates a feasible solution for smart logistics network planning optimization; In sub-steps 1-3, the plant community calculates the optimal solution for the multi-objective optimization of the smart logistics network planning on the smart logistics edge device; the smart logistics edge device sets the multiple smart logistics network planning objectives obtained in sub-step 1-1 as the fitness function, inputs it to the plant community, and sets the feasible solution for the smart logistics network planning optimization randomly generated in sub-step 1-2 as the initial plant individual.

3. The method according to claim 2, characterized in that In sub-step 1-1, the smart logistics network planning optimization problem is to solve the optimal set of nodes and edges of the smart logistics network; smart logistics edge devices refer to devices with computing, communication and storage capabilities used in smart logistics operations, including drones, drone remote controls, logistics robots, electric vehicles, mobile computers, portable communication devices, servers, databases, and various types of smart mobile terminals; smart logistics network planning goals include smart logistics network goals, smart logistics transportation tool goals, and smart logistics operation performance goals; smart logistics network goals include the layout of logistics nodes, the connection of logistics edges, the total length of routes or edges, and qualitative and quantitative descriptions of the number of logistics nodes, the distance between logistics nodes, storage capacity, and storage costs; Smart logistics transportation tool targets include the number of transportation tools, types of transportation tools, transportation tool costs, transportation tool speeds, and transportation tool capacities; Smart logistics operation performance targets include service quality level, carbon emission level, order completion rate, and customer satisfaction. Logistics managers determine computing tasks based on the type and scale of smart logistics and select multiple smart logistics network planning targets. In sub-step 1-2, after the logistics manager selects multiple smart logistics network planning targets on the smart logistics edge device, they are transmitted to multiple smart logistics edge devices through distributed communication; each smart logistics edge device that completes the distributed communication randomly generates a feasible solution for smart logistics network planning optimization according to multiple smart logistics network planning targets; further, each smart logistics edge device calculates a feasible solution for multi-objective optimization of smart logistics network planning, which corresponds to an individual plant, and the smart logistics edge devices participating in the multi-objective optimization of smart logistics network planning jointly compare the fitness of feasible solutions through distributed communication, and jointly select the optimal solution for multi-objective optimization of smart logistics network planning with the highest fitness; In sub-steps 1-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics network planning process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics network planning optimization, that is, encoding and solving the set of nodes and edges of the smart logistics network; setting the growth variation probability, flowering probability, and fruiting probability of the plant community; initializing the smart logistics network set of the individual plant, clearing the plant sowing set, clearing the plant community flowering set, clearing the plant community result pairing set, and clearing the set with the highest ranking priority. The first level of plant individuals; further, the starting and ending conditions of the plant community algorithm are initialized, including the calculation start time, the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; in the subsequent iterations, different from the first iteration, the plant community generates a multi-objective feasible solution for the smart logistics network planning on the edge devices in the smart logistics network planning process, calculates the fitness function value of the plant individuals, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individuals with the best fitness; When the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for multi-objective optimization of smart logistics network planning, that is, the optimal set of smart logistics network nodes and edges.

4. The method according to claim 1, characterized in that In step 2, the following sub-steps are included: Sub-step 2-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics network planning; specifically: the optimal solution of the multi-objective optimization of the smart logistics network planning obtained in sub-steps 1-3 is broadcast to the entire network, and the edge devices receive the broadcast during the smart logistics planning process, and the optimal solution of the multi-objective optimization of the smart logistics network planning is stored in this node, and the multi-objective feasible solutions of the smart logistics network planning previously calculated and stored are cleared; The smart logistics edge device constrains the plant community search based on the optimal solution of the multi-objective optimization of the smart logistics network planning, that is, the plant community searches for a feasible solution for the optimization of the smart logistics plan in the set of nodes and edges of the optimal smart logistics network; each smart logistics edge device calculates a feasible solution for the multi-objective optimization of the smart logistics plan, which corresponds to an individual plant, and the smart logistics edge devices participating in the multi-objective optimization of the smart logistics plan jointly prepare to solve the optimization of the smart logistics plan; In sub-step 2-2, the logistics planner inputs multiple smart logistics planning targets on the smart logistics edge device. Specifically, the logistics planner refers to the optimal solution of multi-objective optimization of smart logistics network planning, inputs the real smart logistics planning targets through the smart logistics edge device, and converts them into smart logistics plans in the virtual world. In sub-step 2-3, the plant community calculates the optimal solution of the multi-objective optimization of the smart logistics plan on the smart logistics edge device; specifically: the smart logistics edge device uses the multi-objective optimization optimal solution of the smart logistics network planning obtained in sub-step 2-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics plan goals obtained in sub-step 2-2 are used as the fitness function of the plant community, and input into the plant community.

5. The method according to claim 4, characterized in that In sub-step 2-1, the smart logistics plan optimization problem is to solve the optimal smart logistics plan schedule and transportation tool set; the smart logistics edge device where the plant individual with the highest fitness is located; In sub-step 2-2, multiple smart logistics planning goals are included; the smart logistics planning goals include the start time and expected deadline of the smart logistics plan, the name of the shipper, the name of the consignee, the shipping address, the receiving address, the code and name of the goods, the type of goods, the quantity of goods, the weight of goods, the distribution network node set, the total number of customers, the set of road sections between nodes, the total number of transport tools, the type of transport tools, the weight limit of transport tools, the transportation mileage, customer customization requirements and additional requests, the time of arrival at the transit logistics node, the time of departure from the transit logistics node, the earliest time for the customer to receive the service, the latest time for the customer to receive the service, the unit time penalty cost of the transport tool for arriving at the customer point in advance, the unit time penalty cost of the transport tool for arriving at the customer point late, the cost of the transport tool traveling a unit distance, the fixed usage cost of the transport tool, and the cost of the transport tool at Real-time load between customer points, speed of transport vehicles between customer points, fuel consumption of transport vehicles, fuel consumption factor, deadweight of transport vehicles, fuel consumption rate when transport vehicles are empty, fuel consumption rate when transport vehicles are fully loaded; transport vehicles refer to various types of logistics vehicles used in smart logistics, including drones, unmanned vehicles, robots, fuel vehicles, electric vehicles, trains, airplanes, and ships; smart logistics planning goals can measure carbon emissions from a macro perspective of a country, region, or industry, conduct statistical analysis on different energy consumption in a country, region, or industry, and infer carbon dioxide emissions based on fuel consumption in statistical data for a certain period of time; multiply the consumption of different fuels in the statistical period by their corresponding fuel consumption factors; further count, classify or summarize different transport vehicles and fuel types to obtain the total amount of carbon dioxide emissions in the statistical period; In sub-step 2-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics planning process, the numerical type and numerical size of the plant individual, and encoding the plant individual according to the optimization of the smart logistics plan, that is, encoding and solving the smart logistics plan schedule and the set of transportation tools; each plant individual corresponds to a feasible solution for the multi-objective optimization of the smart logistics plan, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; Initialize the smart logistics plan set of individual plants, clear the plant sowing set, clear the plant community flowering set, clear the plant community result pairing set, and clear the plant individuals with the highest sorting priority; Furthermore, the starting and ending conditions of the plant community algorithm are initialized, including the calculation start time, the calculation end time or the limit of the number of iterative calculations, and the end error judgment threshold; in the subsequent iterations, different from the first iteration, the plant community generates a multi-objective feasible solution of the smart logistics plan on the edge devices of the smart logistics planning process, calculates the fitness function value of the individual plant, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individual with the best fitness; When the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics plan, that is, the optimal smart logistics plan schedule and transportation tool set.

6. The method according to claim 1, characterized in that In step 3, the following sub-steps are included: Sub-step 3-1, the smart logistics edge device inputs the optimal solution of the multi-objective optimization of the smart logistics plan; specifically: the smart logistics execution optimization problem is to search for the optimal smart logistics execution time period and transportation tool set; in sub-step 2-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution of the multi-objective optimization of the smart logistics plan obtained in sub-step 2-3 to the entire network. During the execution of smart logistics, all edge devices receive the broadcast and store the optimal solution of the multi-objective optimization of the smart logistics plan in this node, and clear the feasible solutions of the multi-objective optimization of the smart logistics plan that were originally calculated and stored; In sub-step 3-2, the logistics executor inputs multiple smart logistics execution targets on the smart logistics edge device. Specifically, the logistics executor refers to the optimal solution of the multi-objective optimization of the smart logistics plan, inputs the actual smart logistics execution targets through the smart logistics edge device, and converts them into smart logistics execution in the real world. In sub-step 3-3, the plant community calculates the optimal solution for multi-objective optimization of smart logistics execution on the smart logistics edge device; specifically: the smart logistics edge device uses the multi-objective optimization optimal solution of the smart logistics plan obtained in sub-step 3-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics execution goals obtained in sub-step 3-2 as the fitness function of the plant community, and inputs them into the plant community.

7. The method according to claim 6, characterized in that In sub-step 3-1, the smart logistics edge device constrains the plant community search according to the optimal solution of the smart logistics plan multi-objective optimization, that is, the plant community searches for a feasible solution for the smart logistics execution optimization in the optimal smart logistics plan schedule and transportation tool set; each smart logistics edge device calculates a feasible solution for the smart logistics execution multi-objective optimization, which corresponds to an individual plant, and the smart logistics edge devices participating in the smart logistics execution multi-objective optimization jointly prepare to solve the smart logistics execution optimization; In sub-step 3-2, multiple smart logistics execution targets are included; the smart logistics execution targets include the name of the consignor, the name of the consignee, the shipping address, the receiving address, the code and name of the goods, the type of goods, the quantity of goods, the weight of the goods, the logistics nodes passed during the transportation process, the different transportation sections divided according to the different logistics nodes passed, each transportation section uses different transportation tools, the name and code of the logistics executor responsible for each transportation section, the start time and end time of each transportation section, the signature of each transportation section, and the intact status of the goods in each transportation section; In sub-step 3-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics execution process, the numerical type and numerical size of the plant individual, and encoding the plant individual according to the smart logistics execution optimization, that is, encoding into a smart logistics execution time period and a set of transportation tools; each plant individual corresponds to a smart logistics execution multi-objective optimization feasible solution, and sets the growth variation probability, flowering probability, and fruiting probability of the plant community; Initialize the smart logistics execution set of individual plants, clear the plant sowing set, clear the plant community flowering set, clear the plant community result pairing set, and clear the plant individuals with the highest sorting priority; further, initialize the starting and ending conditions of the plant community algorithm, including the calculation start time, calculation end time or iteration calculation number limit, and end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates smart logistics execution multi-objective feasible solutions on the edge devices of the smart logistics execution process, calculates the fitness function value of the individual plant, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individuals with the best fitness; When the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for multi-objective optimization of smart logistics execution, that is, the optimal smart logistics execution time period and transportation tool set.

8. The method according to claim 1, in step 4, comprises the following steps: Sub-step 4-1, the smart logistics edge device inputs the optimal solution of smart logistics execution multi-objective optimization; the smart logistics reverse network optimization problem, that is, solving the optimal set of reverse goods and reverse transportation routes; In sub-step 3-3, the smart logistics edge device where the plant individual with the highest fitness is located broadcasts the optimal solution for multi-objective optimization of smart logistics obtained in sub-step 3-3 to the entire network. All edge devices in the smart logistics reverse network receive the broadcast and store the optimal solution for multi-objective optimization of smart logistics in this node, while clearing the feasible solutions for multi-objective optimization of smart logistics that were originally calculated and stored. In sub-step 4-2, the logistics reverse personnel input multiple smart logistics reverse network optimization goals on the smart logistics edge device; specifically: the logistics reverse personnel refer to the smart logistics execution multi-objective optimization optimal solution, input the actual smart logistics reverse goals through the smart logistics edge device, and convert them into the real-world smart logistics reverse operation; In sub-step 4-3, the plant community calculates the optimal solution for multi-objective optimization of the smart logistics reverse network on the smart logistics edge device; specifically: the smart logistics edge device uses the optimal solution for multi-objective optimization of smart logistics obtained in sub-step 4-1 as the parameter that the plant community solution process depends on, and the multiple smart logistics reverse objectives obtained in sub-step 4-2 as the fitness function of the plant community, and inputs them into the plant community.

9. The method according to claim 8, characterized in that In step 4-1, the smart logistics edge device constrains the plant community search according to the optimal solution of the smart logistics execution multi-objective optimization, that is, the plant community searches for a feasible solution of the smart logistics reverse optimization in the optimal smart logistics execution time period and transportation tool set; each smart logistics edge device calculates a feasible solution of the smart logistics reverse network multi-objective optimization, that is, corresponding to an individual plant, and the smart logistics edge devices participating in the smart logistics reverse network multi-objective optimization jointly prepare to solve the smart logistics reverse network optimization; In step 4-2, multiple smart logistics reverse targets are included; the smart logistics reverse targets include the reason for return or exchange, the name of the returner, the name of the consignee, the address of the returner, the consignee address, the code and name of the returned goods, the intact state of the returned goods, the type of returned goods, the quantity of returned goods, the weight of the returned goods, the logistics nodes passed through during the reverse transportation process, the reverse transportation sections divided according to the different reverse logistics nodes passed through, each reverse transportation section using different reverse transportation tools, the name and code of the logistics reverse personnel responsible for each reverse transportation section, the start time and end time of each reverse transportation section, the signature of each reverse transportation section, and the intact state of the goods in each reverse transportation section; In step 4-3, in the first iterative calculation, the plant community performs parameter initialization, including setting the size of the plant population in the plant community according to the number of edge devices in the smart logistics reverse process, the numerical type and numerical size of the individual plant, and encoding the individual plant according to the smart logistics reverse optimization, that is, encoding into a set of reverse goods and reverse transportation routes; each individual plant corresponds to a feasible solution for the multi-objective optimization of the smart logistics reverse network, setting the growth variation probability, flowering probability, and fruiting probability of the plant community; initializing the smart logistics reverse set of individual plant plants, clearing the plant sowing set, clearing the plant community flowering set, and clearing the plant community result pairing set. Combine, clear the plant individuals with the highest sorting priority; further, initialize the starting and ending conditions of the plant community algorithm, including the calculation start time, calculation end time or iteration calculation number limit, and end error judgment threshold; in subsequent iterations, different from the first iteration, the plant community generates a multi-objective feasible solution of the smart logistics reverse network on the edge devices of the smart logistics reverse process, calculates the fitness function value of the plant individuals, and different smart logistics edge devices compare the fitness function values ​​with each other through distributed communication. Through continuous sowing, growth, flowering, and fruiting, all smart logistics edge devices jointly select the plant individuals with the best fitness; When the number of iterations meets the requirements or the iteration error meets the requirements, all smart logistics edge devices select the plant individual with the highest fitness as the optimal solution for the multi-objective optimization of the smart logistics reverse network, that is, the optimal set of reverse cargo and reverse transportation routes.

Images