Cigarette distribution path optimization model and optimization method thereof, electronic equipment and storage medium

Abstract

The invention discloses a cigarette distribution path optimization model and an optimization method thereof, electronic equipment and a storage medium, and the method comprises the steps: firstly, building a cigarette distribution path optimization model which is a robust optimization-based cigarette distribution path optimization model; the distribution cost of the cigarette distribution path optimization model comprises fixed cost, distribution time cost, quantity loss cost and punishment cost, the optimization target is to minimize the sum of the fixed cost, the distribution time cost, the quantity loss cost and the punishment cost, and vehicle routes, time windows and capacity constraints are comprehensively considered. A robust optimization theory is introduced to process uncertainty of requirements, global search and local optimization are balanced by integrating multiple intelligent optimization algorithms and adopting a multi-strategy integration method, and solution quality can be guaranteed while solution efficiency is improved. According to the method, the distribution scheme with higher robustness can be generated under the condition that the demand randomly fluctuates, and the stability and the reliability of the distribution plan in practical application are ensured by defining the uncertainty set of the demand and optimizing the cost under the worst condition, so that the plan failure risk caused by the demand fluctuation is reduced.

Description

Technical Field

[0001] The present invention relates to the field of supply chain logistics distribution, and particularly relates to a cigarette distribution route optimization model, an optimization method, an electronic device, and a storage medium thereof. Background Art

[0002] In the field of cigarette distribution, the optimization of the distribution route is a key issue. Traditional route optimization methods usually assume that the demand is deterministic. However, in reality, the demand for cigarettes (including product specifications and quantities) from each commercial company often has uncertainty. This uncertainty may lead to the failure of the distribution plan, increase the distribution cost, and even affect the service quality. In the prior art, although there are some methods attempting to solve the problem of demand uncertainty, they often have problems such as complex calculation and low efficiency, and are difficult to be popularized in practical applications.

[0003] For example, traditional vehicle routing problems (VRP) and their variants, such as vehicle routing problems with time windows (VRPTW), usually assume that the demand is known. However, in cigarette distribution, due to the randomness of the demand at the receiving points, these methods are difficult to be directly applied. In recent years, some studies have begun to focus on the route optimization problem under demand uncertainty, but most methods either rely on complex mathematical programming with high computational costs or are inefficient in solving large-scale problems. In addition, cigarette distribution also needs to consider soft time window constraints, that is, a certain degree of early or late arrival is allowed, but corresponding waiting or penalty costs will be generated, and there is a lack of research on the combined impact of soft time windows and demand uncertainty in the prior art.

[0004] Therefore, it is necessary to propose an improved cigarette distribution route optimization method that can effectively handle demand uncertainty and take into account soft time window constraints to improve the distribution efficiency, reduce costs, and improve service quality. Summary of the Invention

[0005] In view of this, an embodiment of the present invention provides a cigarette distribution route optimization model and an optimization method thereof. By constructing an optimization model for the cigarette distribution route and considering the uncertainty of demand, a hybrid variable neighborhood search and particle swarm optimization algorithm is used to achieve an efficient solution to the distribution route.

[0006] The technical solution adopted by the present invention to solve the above technical problems is as follows: A cigarette distribution route optimization model, where the model is a cigarette distribution route optimization model based on robust optimization; Define the graph G = (N, E) to represent the considered distribution network, and respectively represent the node set and path set in the graph; represents the set of shipping warehouses, represents the set of receiving points, , represents the set of delivery vehicles; the path between locations and is represented as ; the set of paths in the network can be represented as ; for vehicle from location to location the time is represented by , where and ; Let , associated with all and delivery routes: if vehicle travels on path , then , otherwise 0; Let , associated with all and deliveries: if , then the receiving point is served by vehicle , otherwise 0; Let and represent the arrival time and departure time of vehicle at receiving point respectively; is the time window specified for receiving point , where and represent the earliest and latest delivery times respectively; The delivery cost of the cigarette delivery path optimization model includes fixed cost, delivery time cost, quantity loss cost and penalty cost; The fixed cost includes various types of costs for using vehicles for delivery tasks, including vehicle repair and maintenance costs, driver labor costs, and vehicle depreciation costs; the fixed cost is expressed as:

[0007] where, represents the fixed cost of each vehicle used, represents that vehicle starts from the shipping warehouse o (i.e., the vehicle is used), otherwise 0; The delivery time cost includes variable costs related to the driving route and service receiving points during delivery, and is the product of the total delivery time of the vehicle and the cost per unit time of the corresponding vehicle. Assuming that the speed may vary depending on the vehicle and the time on the same route may be different, the delivery cost is expressed as:

[0008] where, represents the waiting time of vehicle at the receiving point ; The quantity loss cost may be caused by factors such as breakage. Assuming the loss rate is , representing the loss ratio of cigarettes during transportation, the quantity loss cost is expressed as:

[0009] where, represents the unit cost of cigarettes; represents the demand quantity (number of cigarettes) at the receiving point ; The penalty cost is related to the vehicle arriving at the receiving point early or late. The penalty cost is determined according to the arrival time and can be expressed as:

[0010] where, and are the unit waiting and delay penalty costs respectively; Assume that the random demand variable varies within the symmetric interval , where and represent the demand mean and the maximum deviation respectively. Define the auxiliary variable , which varies within [-1, +1]; the budget uncertainty set is ; The objective of the cigarette delivery route optimization model is: .

[0011] As an improvement of the above technical solution, the cigarette delivery route optimization model is subject to a series of constraint conditions, including: Vehicle route constraint: Ensure that the vehicle departs from the shipping warehouse and returns; Capacity constraint: Ensure that the load of each vehicle does not exceed the maximum capacity; Time window constraint: Ensure that the arrival and departure times of the vehicle meet the time window requirements; Decision variable constraint: Ensure that the decision variable is binary.

[0012] The present invention adopts the following technical solutions to solve its technical problems: An optimization method for a cigarette distribution route optimization model, including S1 Construct an initial particle swarm S1.1 Adopt an improved natural integer numbering method to generate an initial solution through heuristic rules, and randomly select a combination of a shipping warehouse and a receiving point; S1.2 Use a greedy strategy to allocate vehicles according to the demand and time window constraints of the receiving points to ensure that the initial solution satisfies the capacity constraint and time window constraints; The position of each particle represents a distribution plan, and the initial velocity is randomly generated. The initial solution considers vehicle routes and time window restrictions to generate a relatively balanced initial particle swarm; S2 Global search based on particle swarm optimization (PSO) S2.1 Adopt a dynamic inertia weight to update the particle velocity and position. The formula is:

[0013]

[0014] where and are adaptive acceleration factors, and are random numbers, w(t) is the dynamic inertia weight, with a large initial value that gradually decreases linearly to the minimum; S2.2 Use the penalty function method to handle constraints Calculate the particle fitness according to the fitness function; when the particle solution violates vehicle routes, time windows, or capacity constraints, increase the fitness penalty to guide the particle to avoid invalid solutions; S2.3 Repeat the operation until the particle swarm optimization termination condition is reached, that is, when the algorithm reaches the preset maximum number of iterations (such as 1000 times), or when the global optimal solution gBest has not improved in 50 consecutive iterations, stop the search, update the particle position, and record the local optimal and global optimal ; S3 Local optimization based on variable neighborhood search (VNS) S3.1 On the basis of S2, select the current solution of the particle and iteratively apply the following neighborhood search operations: S3.1.1 Repositioning: Select a delivery point to remove and reassign it to another vehicle to optimize driving and waiting times; S3.1.2 Exchange: Select two vehicles to exchange service points and adjust the load distribution; S3.1.3 Reversal: Select a path to reverse the order to optimize the total time; S3.2 Evaluate the fitness of the current solution of the particle and update the local optimum ; S3.3 Repeat the operation until the improvement of local optimality is no longer significant; S4 Global search based on the simulated annealing SA strategy S4.1 Initialize the temperature and the cooling rate

[0015] Set the initial temperature , which decreases with the number of iterations , where ; A high initial temperature allows a large search space, and a low temperature in the later stage focuses on local optimization; S4.2 Decide whether to accept a worse solution according to the acceptance probability of simulated annealing When generating a new solution, if the fitness is worse , decide whether to accept the worse solution according to the acceptance probability ; S4.3 Decrease the temperature T and judge whether the termination temperature or the convergence condition is reached. When the temperature T drops to , or when the global optimal solution has not improved in 100 consecutive iterations, stop the search; S5 Improve the solution diversity based on the genetic algorithm GA S5.1 Adopt the path encoding method, and each chromosome represents a complete delivery plan; The chromosome consists of a series of sequences of delivery points and vehicle allocations. The sequence of delivery points is , and the sequence of vehicle allocations is ; S5.2 Initialize the population of the genetic algorithm The initial population consists of the excellent solutions of PSO and randomly generated solutions to ensure the diversity of the population; Convert the and gBest solutions of PSO into the chromosome form of GA, put them into the initial population, randomly assign delivery points to vehicles, check the capacity constraint , and randomly sort the sequence of delivery points for each vehicle to generate a path; S5.3 Perform the selection operation through the roulette wheel selection and elitist retention strategies The roulette wheel selection method is adopted to select parental individuals based on fitness proportion; The specific probability is as follows: wherein, is the fitness of individual i, and N is the population size; The elite retention strategy is adopted to directly retain the optimal individual to the next generation to avoid the loss of excellent solutions; In S5.4, the crossover operation adopts partially matched crossover (PMX) or ordered crossover (OX) to ensure that the offspring meet the constraints; S6.5 Mutation operation The mutation operation calculates the population diversity index based on adaptive parameters : If D > 0.7, set the crossover rate = 0.8, and the mutation rate = 0.1; If D < 0.3, set the crossover rate = 0.6, and the mutation rate = 0.2; Dynamic adjustment formula: , ; wherein, , , , ; S6.6 Evaluate the robustness of the new solution under demand uncertainty Introduce the simulated annealing acceptance mechanism to perturb the demand Evaluate the worst-case scenario cost , and retain the solution with the best robustness; S7 Termination condition judgment S7.1 When the number of iterations reaches the predetermined value or the change in fitness is less than the threshold, stop the algorithm; S7.2 Output the global optimal solution as the final distribution plan.

[0016] The optimization method of this cigarette distribution route optimization model adopts multi-strategy collaborative optimization: the fusion optimization of PSO and GA, VNS, and SA, enhancing the global search ability and local optimization ability: 1) Co-evolution of PSO and GA: The pBest and gBest of PSO are used as the initial population of GA. The crossover operation of GA provides a new direction for PSO, and the crossover operation of GA provides a new search direction for PSO. Specifically, the excellent solutions (pBest and gBest) in PSO are introduced into the initial population of GA to enhance the diversity of the population; 2) VNS is used as a local search strategy in each generation of GA evolution to further optimize the excellent individuals. In each generation of the population generated by GA, the individuals with the top 10% fitness rankings are selected, and the relocation, exchange, and inversion operations of VNS are applied to further optimize the path and vehicle allocation; 3) In the mutation and crossover operations of GA, the acceptance probability of SA is introduced to allow the acceptance of worse solutions under certain conditions to jump out of the local optimum.

[0017] Meanwhile, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the optimization method of the cigarette distribution path optimization model as described above are implemented.

[0018] Meanwhile, the present invention also proposes a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program is executed by the processor, the steps of the optimization method of the cigarette distribution path optimization model as described above are implemented.

[0019] Beneficial effects brought by the present invention: By introducing the Robust Optimization (RO) theory, the present invention constructs a distribution path optimization model considering the uncertainty of cigarette demand. Compared with the limitation of the traditional method assuming deterministic demand, this model can generate a more robust distribution plan under the condition of random demand fluctuations. By defining the uncertainty set of demand and optimizing the cost in the worst case, the stability and reliability of the distribution plan in practical applications are ensured, thus reducing the risk of plan failure caused by demand fluctuations.

[0020] The hybrid optimization algorithm proposed by the present invention combines the advantages of Particle Swarm Optimization (PSO), Variable Neighborhood Search (VNS), Simulated Annealing (SA), and Genetic Algorithm (GA), overcoming the problems of complex calculation and low efficiency of traditional methods. The adaptive acceleration factor, dynamic inertia weight, and diversity-driven parameter adjustment mechanism enable the algorithm to achieve a balance between global search and local optimization. Compared with the existing technologies relying on complex mathematical programming, this method shows higher computational efficiency and practicability when solving large-scale problems.

[0021] The present invention generates an initial solution through an improved natural integer numbering method and heuristic rules, and combines multi-strategy fusion (such as the co-evolution of PSO and GA, the local optimization of VNS, and the jumping-out mechanism of SA), significantly improving the quality of the initial solution and the convergence speed of the algorithm. This method effectively avoids the defect that traditional optimization algorithms are prone to fall into local optima, ensuring the global optimality of the solution results. Description of the Drawings

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0023] Figure 1 It is a schematic flowchart of the optimization method provided by the embodiment of the present invention; Figure 2 It is a schematic structural diagram of an electronic device provided by the embodiment of the present invention. Detailed implementation manners

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0025] Moreover, the following description is for illustration rather than limitation. Specific details such as specific system structures and technologies are proposed to thoroughly understand the embodiments of the present invention. However, those skilled in the art should clearly understand that the present invention can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present invention.

[0026] The first embodiment of the present invention relates to a cigarette delivery path optimization model, and the model is a Cigarette Delivery Path Optimization with Time Windows (CDPPPTW) model based on robust optimization.

[0027] The distribution network considered in this embodiment consists of one or more shipping warehouses and multiple receiving points with known geographical locations. The shipping warehouses have enough vehicles to perform the distribution tasks. These vehicles depart from the shipping warehouses and distribute cigarettes to all receiving points according to the planned routes.

[0028] These distribution vehicles have different maximum load capacities and may also have different driving speeds. Each distribution vehicle can serve multiple receiving points. Once the vehicle arrives at the receiving point, they must provide an unloading service. After the service is completed, the vehicle returns to the shipping warehouse.

[0029] Assume that each delivery destination has a known time window; however, the time windows may vary among different delivery destinations. A vehicle may arrive within the time window outside the delivery destination; however, if the vehicle arrives at the delivery destination earlier, it must wait until the earliest delivery service time. Therefore, an additional waiting cost will be incurred. On the contrary, if the vehicle arrives at the delivery destination late, an overtime penalty cost will be incurred. The demand (product specifications and quantity) for cigarettes of each commercial company is not fixed, and all demands must be met. The objective of this model is to generate a delivery route solution with the lowest delivery cost.

[0030] For this purpose, this embodiment defines a graph G = (N, E) to represent the considered delivery network, and represent the set of nodes and the set of paths in the graph respectively; represents the set of shipping warehouses, represents the set of delivery destinations, , represents the set of delivery vehicles; the path between locations and is denoted as ; the set of paths in the network can be denoted as ; the time for vehicle to travel from location to location is represented by , where and ; Let , associated with all and for the delivery route: if vehicle travels on path , then , otherwise 0; Let , associated with all and for the delivery: if , then the delivery destination is served by vehicle , otherwise 0; Let and represent the arrival time and departure time of vehicle at the delivery destination respectively; the symbol and variable definitions of a more detailed model can be referred to in Table 1 below.

[0031] Table 1 Model Symbols and Variables

[0032] The distribution costs of the CDPPPTW model include fixed costs, distribution time costs, quantity loss costs, and penalty costs.

[0033] The fixed costs include various types of costs for using vehicles for distribution tasks, including vehicle repair and maintenance costs, driver labor costs, and vehicle depreciation costs; the fixed costs are expressed as:

[0034] where, represents the vehicle departing from the shipping warehouse o (i.e., the vehicle is in use), otherwise 0.

[0035] The distribution time costs include variable costs related to the routes traveled and service receiving points during the distribution process, which is the product of the total distribution time of the vehicle and the cost per unit time of the corresponding vehicle; assuming that the speed may vary by vehicle and the time of the vehicle on the same route may be different, the distribution costs are expressed as:

[0036] where, represents the vehicle at the receiving point waiting time.

[0037] The quantity loss costs may be caused by factors such as breakage. Assuming the loss rate is , which represents the loss ratio of cigarettes during transportation, the quantity loss costs are expressed as: .

[0038] The penalty costs are related to the vehicle arriving at the receiving point early or late. Define a penalty cost function , and determine the penalty costs according to the arrival time , which can be expressed as:

[0039] where, and are the unit waiting and delay penalty costs respectively.

[0040] Considering the uncertainty of cigarette demand, this embodiment adopts the robust optimization (RO) theory to formulate the problem. Assume that the random demand variable varies within the symmetric interval , where and represent the demand mean and the maximum deviation respectively; define the auxiliary variable , which varies within [-1, +1]; define the budget uncertainty set and the robust optimization objective: The objective of the CDPPPTW model is:

[0041] Among them, demand uncertainty is handled through constraints.

[0042] The CDPPPTW model is restricted by a series of constraint conditions, including: Vehicle route constraint: Ensure that the vehicle departs from the shipping warehouse and returns. Capacity constraint: Ensure that the load of each vehicle does not exceed the maximum capacity. Time window constraint: Ensure that the arrival and departure times of the vehicle meet the time window requirements. Decision variable constraint: Ensure that the decision variable is binary.

[0043] The second embodiment of the present invention relates to an optimization method for a cigarette distribution route optimization model, that is, a hybrid optimization method for solving the CDPPPTW model of the first embodiment. By integrating multiple intelligent optimization algorithms and combining the robust optimization theory, it efficiently solves the distribution route planning problem with time window and capacity constraints.

[0044] Refer to Figure 1 , the optimization method of this embodiment includes the following steps: S1 Construct an initial particle swarm S1.1 Adopt an improved natural integer numbering method, generate an initial solution through heuristic rules, and randomly select a combination of the shipping warehouse and the receiving point. S1.2 Use a greedy strategy to allocate vehicles according to the demand and time window constraints of the receiving points to ensure that the initial solution meets the capacity constraint and the time window constraint; S1.3 The position of each particle represents a distribution plan, and the initial velocity is randomly generated. The initial solution considers the vehicle route and time window restrictions to generate a relatively balanced initial particle swarm; In this step, when initializing the position of the particle, in addition to considering the allocation of vehicles, the route and time window restrictions of each vehicle are also considered. In this way, a relatively balanced initial solution can be generated to avoid getting stuck at the local optimal solution.

[0045] S2 Global search based on particle swarm optimization PSO The PSO algorithm is used for global search and can explore the solution space within a large range. To further improve the search ability of the PSO algorithm, the following improvements are made in this step: S2.1 Adopt a dynamic inertia weight Update the particle velocity and position, and the formula is:

[0046]

[0047] Among them, and are adaptive acceleration factors. As the iteration process progresses, the acceleration factors are gradually adjusted. At the beginning, the acceleration factors are larger, allowing the particles to explore a wider area within the solution space; as the algorithm progresses, the acceleration factors gradually decrease, and the particles tend to local search, thus avoiding premature convergence to local optima; and are random numbers, w(t) is the dynamic inertia weight, with a large initial value that gradually decreases linearly to the minimum, encouraging the particles to rely more on the previous information during the later search process to enhance the balance between global search and local search; S2.2 Use the penalty function method to handle constraints During the update process of PSO, the penalty function method is used to handle vehicle routes, time windows, and capacity constraints.

[0048] Calculate the particle fitness according to the fitness function; when the solution of the particle violates the constraints, the corresponding fitness will increase, prompting the particle to avoid these invalid solutions; S2.3 Repeat the operation until the particle swarm optimization termination condition is reached, update the particle position, and record the local optimum and the global optimum .

[0049] S3 Local optimization based on variable neighborhood search VNS Based on particle swarm optimization, this step adopts VNS as the local optimization strategy to further improve the solution quality. VNS enhances the local optimality of the solution by iteratively applying different neighborhood search operations.

[0050] S3.1 Based on S2, select the current solution of the particle and iteratively apply the following neighborhood search operations: S3.1.1 Re-location: Select a delivery point to remove and re-allocate it to another vehicle to optimize the driving and waiting times; S3.1.2 Exchange: Select two vehicles to exchange service points to adjust the load distribution; S3.1.3 Inversion: Select a path to reverse the order to optimize the total time; S3.2 Evaluate the fitness of the current solution of the particle and update the local optimum ; S3.3 Repeat the operation until the improvement of local optimality is no longer significant.

[0051] S4 Global search based on the simulated annealing SA strategy To further improve the quality of the solution and avoid being trapped in local optima, this step adds the simulated annealing (SA) strategy on the basis of VNS. Simulated annealing enhances the global search ability by allowing occasional acceptance of worse solutions during the search process to jump out of local optima.

[0052] S4.1 Initialize the temperature and the cooling rate

[0053] Set the initial temperature , which decreases with the number of iterations , where ; A high initial temperature allows a large search space, and a low temperature in the later stage focuses on local optimization; S4.2 Determine whether to accept a worse solution according to the acceptance probability of simulated annealing When generating a new solution, if the fitness is worse , decide whether to accept the worse solution according to the acceptance probability ; S4.3 Decrease the temperature T and judge whether the termination temperature or convergence condition is reached.

[0054] S5 Improve the solution diversity based on the genetic algorithm GA This step designs an adaptive parameter adjustment mechanism to dynamically adjust parameters according to the diversity and fitness distribution of the population to balance exploration and exploitation.

[0055] S5.1 Adopt the path encoding method, where each chromosome represents a complete delivery plan; The chromosome consists of a series of sequences of delivery points and vehicle allocations. The sequence of delivery points is , and the sequence of vehicle allocations is ; S5.2 Initialize the population of the genetic algorithm The initial population consists of the excellent solutions of PSO and randomly generated solutions to ensure the diversity of the population; Convert the local optimum and the global optimum gBest solution of PSO into the chromosome form of GA, put them into the initial population, randomly allocate delivery points to vehicles, check the capacity constraint , and randomly sort the sequence of delivery points for each vehicle to generate a path; S5.3 Selection operation is performed through roulette wheel selection and elitist retention strategy The roulette wheel selection method is adopted to select parental individuals based on fitness proportion The specific probability is as follows where is the fitness of individual i, and N is the population size The elitist retention strategy is adopted to directly retain the optimal individual to the next generation, avoiding the loss of excellent solutions In addition, in each generation, individuals with the top 10% fitness rankings are screened out from the current population and forced to be retained in the next generation to strengthen the inheritance of excellent solutions, improve the convergence speed and stability. This operation can be performed after elitist retention to ensure that it is not replaced by roulette wheel selection S5.4 The crossover operation adopts partially mapped crossover (PMX) or order crossover (OX) to ensure that the offspring meet the constraints S6.5 Mutation operation The mutation operation calculates the population diversity index based on adaptive parameters : If D > 0.7, set the crossover rate = 0.8, and the mutation rate = 0.1 If D < 0.3, set = 0.6, = 0.2 Dynamic adjustment formula , ; where , , , ; S6.6 Evaluate the robustness of the new solution under demand uncertainty Introduce the simulated annealing acceptance mechanism to perturb the demand Evaluate the worst-case scenario cost , and retain the solution with the best robustness The purpose of introducing the worst cost in this step is to evaluate and improve the robustness of the distribution plan under demand uncertainty. By analyzing the cost performance of the plan in the worst-case scenario, the algorithm can screen out solutions that can not only maintain a low average cost but also operate stably in extreme situations

[0056] S7 Termination condition judgment S7.1 When the number of iterations reaches a predetermined value or the fitness change is less than the threshold, such as when the number of iterations reaches 500 times, or the fitness improvement is less than 0.1% in 50 consecutive iterations, stop the algorithm The S7.2 output global optimal solution as the final distribution plan.

[0057] Figure 2 It is a schematic diagram of the electronic device 10 provided by another embodiment of the present invention. As Figure 2 shown, the electronic device 10 of this embodiment includes: a processor 11, a memory 12, and a computer program 13 stored in the memory 12 and executable on the processor 11, such as an optimization method program of a cigarette distribution path optimization model. When the processor 11 executes the computer program 13, the steps in the above-mentioned embodiments of the optimization method of each cigarette distribution path optimization model are implemented, such as Figure 1 the steps shown.

[0058] Exemplarily, the computer program 13 can be divided into one or more modules / units. One or more modules / units are stored in the memory 12 and executed by the processor 11 to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program 13 in the electronic device 10.

[0059] The electronic device 10 can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The electronic device 10 may include, but is not limited to, a processor 11 and a memory 12. Those skilled in the art can understand that Figure 2 this is only an example of the electronic device 10, and does not constitute a limitation on the electronic device 10. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, the electronic device 10 may further include input / output devices, network access devices, buses, etc.

[0060] The so-called processor 11 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0061] The memory 12 can be an internal storage unit of the electronic device 10, such as the hard disk or memory of the electronic device 10. The memory 12 can also be an external storage device of the electronic device 10, such as a plug-in hard disk equipped on the electronic device 10, a Smart Media Card (SMC), a Secure Digital (SD) card, a FlashCard, etc. Further, the memory 12 can also include both the internal storage unit of the electronic device 10 and the external storage device. The memory 12 is used to store computer programs and other programs and data required by the electronic device 10. The memory 12 can also be used to temporarily store the data that has been output or will be output.

[0062] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of this application. The specific working process of the units and modules in the above system can refer to the corresponding process in the foregoing method embodiment and will not be elaborated here.

[0063] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0064] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0065] In the embodiments provided by the present invention, it should be understood that the disclosed device / electronic device and method can be implemented in other ways. For example, the device / electronic device embodiments described above are merely illustrative. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical or other forms.

[0066] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0067] In addition, the functional units in various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0068] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above method embodiments of the present invention, it can also be completed by a computer program instructing relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device that can carry computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0069] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. Cigarette distribution route optimization model, characterized in that: The model is a cigarette distribution route optimization model based on robust optimization; Define the graph \(G=(N, E)\) to represent the considered distribution network, and represent the set of nodes and the set of paths in the graph respectively; represents the set of shipping warehouses, represents the set of receiving points, , represents the set of distribution vehicles; The path between locations and is denoted as ; The set of paths in the network can be denoted as ; The time for vehicle to travel from location to location is represented by , where and ; Let , associated with all and delivery: If the vehicle is traveling on the path , then , otherwise 0; Let , associated with all and delivery: If , then the receiving point is served by vehicle , otherwise 0; Let and respectively represent the arrival time and departure time of the vehicle at the pick-up point ; is the time window specified for the pick-up point , where and respectively represent the earliest and latest delivery times; The optimization objective of the cigarette distribution route optimization model is to minimize the sum of fixed costs, distribution time costs, quantity loss costs, and penalty costs; The fixed costs include the fixed costs of using vehicles for distribution tasks, and the fixed costs are expressed as: Among them, represents the fixed cost of each vehicle used, represents the vehicle starting from the shipping warehouse o, otherwise 0; The delivery time cost is the product of the total delivery time of the vehicle and the cost per unit time of the corresponding vehicle and the delivery cost is expressed as: Among them, represents the waiting time of the vehicle at the receiving point ; represents the unloading service time of the vehicle at the receiving point ; The quantity loss cost: Assuming the loss rate is , which represents the loss ratio of cigarettes during transportation, the quantity loss cost is expressed as: Among them, represents the unit cost of cigarettes; represents the receiving point and the demand at that point; The penalty cost is related to the vehicle arriving at the receiving point earlier or later, and the penalty cost is determined according to the arrival time and is expressed as: Among them, and are the unit waiting and delay penalty costs respectively; Assume that the random demand variable varies within the symmetric interval where and represent the mean demand and the maximum deviation respectively; define the auxiliary variable which varies within [-1, +1]; the budget uncertainty set is ; The optimization objective of the cigarette distribution route optimization model is: 。 2. The cigarette distribution route optimization model according to claim 1, characterized in that: The cigarette distribution route optimization model is restricted by a series of constraint conditions, including: Vehicle route constraint: Ensure that the vehicle departs from and returns to the shipping warehouse; Capacity constraint: Ensure that the load of each vehicle does not exceed the maximum capacity; Time window constraint: Ensure that the arrival and departure times of the vehicle meet the time window requirements; Decision variable constraints: Ensure that the decision variables are binary.

3. Optimization method for the cigarette distribution route optimization model, characterized in that: The optimization method is used to solve the cigarette distribution route optimization model according to any one of claims 1 to 2, including S1 Construct an initial particle swarm S1.1 Randomly select combinations of shipping warehouses and receiving points; S1.2 Allocate vehicles according to the demand of receiving points and time window constraints; S1.3 Position of each particle represents a delivery plan, and the initial velocity is randomly generated. The initial solution takes into account vehicle routes and time window constraints to generate an initial particle swarm; S2 Global search based on particle swarm optimization (PSO) S2.1 Adopt a dynamic inertia weight Update the particle velocity and position. The formula is as follows: Among them, and are acceleration factors, and are random numbers, w(t) is the dynamic inertia weight, with a relatively large initial value that gradually decreases to the minimum; S2.2 Use the penalty function method to handle constraints When the particle solution violates vehicle route, time window, or capacity constraints, increase the fitness penalty to guide the particle to avoid invalid solutions; S2.3 Repeat the operation until the particle swarm optimization termination condition is met, update the particle positions and record the local optimum and the global optimum ; S3 Local optimization based on variable neighborhood search (VNS) S3.1 Select the current solution of the particle and iteratively apply the following neighborhood search operations: S3.1.1 Re-location: Select a receiving point to remove and re-allocate it to another vehicle to optimize driving and waiting times; S3.1.2 Exchange: Select two vehicles to exchange service points and adjust the load distribution; S3.1.3 Inversion: Select a path to reverse the order to optimize the total time; S3.2 Evaluate the fitness of the current solution of the particle and update the local optimum ; S3.3 Repeat the operation until the improvement of local optimality is no longer significant; S4 Global search based on simulated annealing (SA) strategy S4.1 Initialize the temperature and the cooling rate Set the initial temperature , decreasing with the number of iterations ; S4.2 Determine whether to accept a worse solution according to the simulated annealing acceptance probability When generating a new solution, if the fitness is poor , determine whether to accept the inferior solution according to the acceptance probability ; S4.3 Reduce the temperature T and judge whether the termination temperature or convergence condition is reached; S5 Improve solution diversity based on genetic algorithm (GA) S5.1 Adopt a path coding method, and each chromosome represents a complete distribution plan; S5.2 Initialize the genetic algorithm population The initial population consists of excellent solutions of PSO and randomly generated solutions; Convert the and gBest solutions of PSO into chromosome form, put them into the initial population, randomly assign delivery points to vehicles, check the capacity constraint, randomly sort the delivery point sequences for each vehicle, and generate routes; S5.3 Perform selection operations through roulette wheel selection and elitist retention strategy; S5.4 Perform partially mapped crossover and / or order crossover operations; S6.5 Mutation operation The mutation operation calculates the population diversity index based on adaptive parameters ; Based on the population diversity index Dynamically adjust the crossover rate and the mutation rate ; S6.6 Evaluate the robustness of the new solution under demand uncertainty Introduce the simulated annealing acceptance mechanism to perturb the demand Evaluate the cost of the worst-case scenario , and retain the solution with the best robustness; S7 Judgment of termination conditions S7.1 When the number of iterations reaches a predetermined value or the change in fitness is less than the threshold, stop the algorithm; S7.2 Output the global optimal solution as the final distribution plan.

4. The optimization method of the cigarette distribution route optimization model according to claim 3, characterized in that: In S2.1, and are adaptive acceleration factors: at the beginning, the acceleration factor is larger, allowing the particles to explore a wider area in the solution space; as the algorithm progresses, the acceleration factor gradually decreases, and the particles tend to local search to avoid premature convergence to local optima; w(t) is a dynamic inertia weight, with a large initial value that gradually decreases linearly to the minimum to enhance the balance between global search and local search.

5. The optimization method of the cigarette distribution route optimization model according to claim 3, characterized in that: In S4.1, set the initial temperature , decreasing with the number of iterations , where .

6. The optimization method of the cigarette distribution route optimization model according to claim 3, characterized in that: In S5.1, a path coding method is adopted, and each chromosome represents a complete distribution plan; The chromosome consists of a series of sequences of delivery points and vehicle allocations. The sequence of delivery points is , and the sequence of vehicle allocations is .

7. The optimization method of the cigarette distribution route optimization model according to claim 3, characterized in that: In S5.3, a selection operation is performed through roulette wheel selection and elitist retention strategy: The roulette wheel selection method is adopted to select parental individuals based on fitness proportion; The specific probability is as follows: Among them, is the fitness of individual i, and N is the population size; The elitist retention strategy is adopted to directly retain the optimal individual to the next generation to avoid the loss of excellent solutions.

8. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 3 to 7 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 3 to 7 are implemented.

Images