Logistics distribution method and system based on Internet of Things

Abstract

The invention relates to the technical field of logistics distribution, in particular to a logistics distribution method and system based on the Internet of Things, and the method comprises the steps: calculating the predicted passing time cost of each road section in a road network in a plurality of time windows in the future through a space-time prediction model; using a Bayesian inference model to generate discrete sign-in success probability distribution for each order in the plurality of time windows; a multi-objective optimization algorithm is applied to solve and obtain a group of Pareto optimal distribution scheme sets; and selecting a final delivery route from the Pareto optimal delivery scheme set according to a preset risk preference or cost benefit strategy, and issuing the final delivery route to a vehicle terminal for execution. According to the invention, scientific basis for selection based on different business strategies can be provided for decision makers, so that the final distribution decision is more flexible and strategic.

Description

Technical Field

[0001] The present invention relates to the field of logistics and distribution technology. More specifically, the present invention relates to a method and system for logistics and distribution based on the Internet of Things. Background Art

[0002] With the rapid development of e-commerce and the instant delivery industry, logistics and delivery efficiency, cost, and customer satisfaction have become key indicators of a company's core competitiveness. To address increasingly complex delivery needs, existing technologies primarily focus on the vehicle routing problem (VRP), aiming to plan a series of optimal customer-visiting routes for multiple vehicles. Traditional logistics and delivery systems typically utilize real-time GPS data and electronic map information, combined with heuristic algorithms (such as genetic algorithms and ant colony algorithms) or precise algorithms to generate delivery plans. These methods address fundamental route planning issues to a certain extent, such as minimizing total mileage or total travel time while satisfying rigid constraints such as vehicle load, volume, and customer time windows. However, these systems still have significant limitations when dealing with the dynamic nature of urban traffic and the uncertainties inherent in the delivery process.

[0003] Existing technologies have shortcomings in several areas: First, when it comes to traffic state prediction, most systems rely on real-time traffic information or simple averages of historical data, making it difficult to accurately predict travel times on road sections within specific time periods. They lack the ability to deeply model the complex spatiotemporal dependencies between road network topology and traffic flows, resulting in planned routes that may be suboptimal in practice due to traffic congestion. Second, when considering delivery success rates, traditional methods typically treat the customer's time window as a fixed constraint, ignoring the probabilistic nature of successful delivery within that window. For example, delivery success rates for different destinations (e.g., office buildings and residential areas) vary significantly over time. Existing systems rarely quantitatively model this uncertainty, resulting in high redelivery rates and increased operating costs. Regarding optimization objectives, most systems employ a single objective or combine multiple objectives into a single objective function through linear weighting. This approach fails to reveal the inherent trade-offs between different objectives (e.g., time cost and delivery success rate), making it difficult to provide flexible and diverse decision options based on different business strategies (e.g., prioritizing cost or service quality). Summary of the Invention

[0004] In order to solve the above technical problem of difficulty in providing flexible and diverse decision-making options according to different business strategies, the present invention provides the following solution.

[0005] A method for logistics distribution based on the Internet of Things, comprising the following steps: obtaining order information to be delivered, real-time status data of vehicles, road network topology structure and historical traffic flow data of the target delivery area; based on the road network topology structure and historical traffic flow data, calculating the predicted travel time cost of each road section in the road network in multiple future time windows by integrating a spatiotemporal prediction model of a graph convolutional network and a gated recurrent unit; based on historical delivery data and time sensitivity, using a Bayesian inference model, generating a discrete probability distribution of successful signing for each order in the multiple time windows; combining the real-time status of the vehicles, calculating the predicted travel time cost of each road section in the road network in multiple future time windows; and The feasibility constraint set for vehicle task execution is constructed based on the endurance, load limit and health status index in the dynamic data; the predicted travel time cost, sign-off success probability distribution and feasibility constraint set are used as input, and a multi-objective optimization algorithm is applied to minimize the total predicted travel time cost of all vehicles and maximize the total sign-off success probability of all orders as independent optimization goals, thereby solving a set of Pareto optimal delivery solutions containing multiple non-inferior solutions; according to the preset risk preference or cost-effectiveness strategy, the final delivery route is selected from the Pareto optimal delivery solution set and sent to the vehicle terminal for execution.

[0006] The present invention can provide decision makers with a scientific basis for making choices based on different business strategies, making the final distribution decision more flexible and strategically valuable.

[0007] Furthermore, the spatiotemporal prediction model integrating the graph convolutional network and the gated recurrent unit is used to calculate the predicted travel time cost of each road section in the road network in multiple future time windows, including: The road network topology is abstracted into a graph, with intersections as nodes and road sections as edges. The spatial dependency features of each road section are extracted using a graph convolutional network. The traffic flow data of each road section within a preset historical period is constructed into a time series and input into a gated recurrent unit to extract the temporal variation pattern. The spatial dependency features are integrated with the temporal variation pattern to output the predicted travel time cost for each time window in the future prediction domain with a preset duration as the unit.

[0008] The present invention makes full use of the characteristics of the road network, constructs a graph model and uses a graph convolutional network for processing.

[0009] Furthermore, the Bayesian inference model is used to generate a discrete probability distribution of successful receipt for each order within the multiple time windows, including: Based on historical delivery data, a priori probability of successful receipt related to the order destination type and delivery time period is constructed; according to the time sensitivity level of the order, a likelihood function is established, where the higher the time sensitivity level of the order, the higher the likelihood value of receipt within the time window required by the order; the Bayesian formula is applied to integrate the prior probability of successful receipt and the likelihood function to obtain the posterior probability of successful receipt of the order in each future time window.

[0010] The present invention uses a Bayesian inference model to quantify the probability of successful order receipt, so that the delivery plan can actively avoid the risk of failed receipt.

[0011] Furthermore, the feasibility constraint set for vehicle mission execution is constructed, including: The cruising range constraint is set to ensure that the total estimated energy consumption of the vehicle's planned route does not exceed the currently available energy; the load limit constraint is set to ensure that the total weight and total volume of the orders carried by the vehicle do not exceed the vehicle's rated load and rated capacity; and the health status constraint is set to ensure that the vehicle's health status index must be higher than the preset threshold.

[0012] Furthermore, the multi-objective optimization algorithm is applied to obtain a set of Pareto optimal distribution solutions containing multiple non-inferior solutions, including: A multi-objective evolutionary algorithm based on non-dominated sorting is used to encode potential delivery plans. The first objective function is set as the sum of the predicted travel time costs of all vehicle routes and is minimized. The second objective function is set as the sum of the successful sign-in probabilities of all orders at their respective delivery time points and is maximized. The Pareto front solution set is generated by iteratively performing selection, crossover and mutation operations.

[0013] The present invention does not directly generate a plan, but generates a set of Pareto optimal delivery plans that reveal the trade-off between time cost and signature success rate, laying the foundation for subsequent steps.

[0014] Furthermore, selecting a final delivery route from the Pareto optimal delivery solution set according to a preset risk preference or cost-effectiveness strategy includes: For each plan in the Pareto optimal delivery plan set, its total predicted travel time cost and total probability of successful signing are normalized; according to the preset risk preference or cost-effectiveness strategy, corresponding weight coefficients are assigned to the normalized total predicted travel time cost and total probability of successful signing; based on the weight coefficients, a weighted comprehensive score of each plan is calculated, and the plan with the best comprehensive score is selected as the final delivery route.

[0015] The present invention provides final delivery routes based on different business strategies, making decision-making more flexible.

[0016] Furthermore, the order information includes the order number, cargo weight, volume, delivery address and customer expected delivery time.

[0017] Furthermore, the multi-objective evolutionary algorithm based on non-dominated sorting is an NSGA-II algorithm.

[0018] Furthermore, the risk preference includes a service priority strategy.

[0019] The present invention also provides a logistics distribution system based on the Internet of Things, including a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the above-mentioned logistics distribution method based on the Internet of Things.

[0020] Compared with the existing technology, the logistics distribution method provided by the present invention has achieved significant beneficial effects by introducing advanced prediction and optimization models. First, it adopts a spatiotemporal prediction model that integrates graph convolutional networks and gated recurrent units, which can more accurately calculate future travel times, provide a reliable data basis for path planning, and improve the actual execution efficiency of the route. Secondly, the present invention uses a Bayesian inference model to quantify the probability of successful receipt of an order, so that the distribution plan can actively avoid the risk of failed receipt, thereby significantly improving the one-time delivery rate and reducing the operating costs caused by secondary distribution. By applying a multi-objective optimization algorithm, the present invention can generate a set of Pareto optimal distribution solutions that reveal the trade-off between time cost and receipt success rate, providing decision makers with a scientific basis for making choices based on different business strategies (such as cost priority or service quality priority), making the final distribution decision more flexible and strategically valuable. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flow chart schematically illustrating a method for logistics distribution based on the Internet of Things according to an embodiment of the present invention; Figure 2 is a schematic diagram schematically illustrating steps S5 and S6 according to an embodiment of the present invention; Figure 3 FIG. 1 is a schematic diagram schematically illustrating a system for logistics distribution based on the Internet of Things according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0023] like Figure 1 As shown, a logistics distribution method based on the Internet of Things includes the following steps: S1, obtains order information to be delivered, real-time status data of vehicles, road network topology of the target delivery area, and historical traffic flow data; Specifically, the order information includes the order number, cargo weight, volume, delivery address and customer expected time; through the on-board Internet of Things terminal, such as the GPS locator and the on-board diagnostic system OBD interface, the vehicle's latitude and longitude coordinates, remaining mileage, current load capacity and equipment health diagnostic code are periodically collected as the real-time status data of the vehicle; by calling the open platform API such as Gaode Map or Baidu Map, the road network topology structure (including a digital road network map consisting of roads and intersections) and historical traffic flow data are obtained (traffic flow data is very rich, including flow, speed and density, etc., and in this embodiment mainly includes the historical average speed or travel time data of the corresponding road section aggregated by hours or minutes in the past few months).

[0024] S2, based on the road network topology and historical traffic flow data, calculates the predicted travel time cost of each road segment in the road network in multiple future time windows by integrating a graph convolutional network with a gated recurrent unit spatiotemporal prediction model; Specifically, the road network topology is abstracted into a graph, with intersections as nodes and road segments as edges. A graph convolutional network is used to extract the spatial dependency characteristics of each road segment. Traffic flow data for each road segment within a preset historical period is constructed into a time series and input into a gated recurrent unit to extract temporal variation patterns. The spatial dependency characteristics are then integrated with the temporal variation patterns to output the predicted travel time for each time window within a preset duration within the future prediction domain, i.e., the predicted travel time cost.

[0025] For example, consider a road network in a central city area, encompassing hundreds of intersections and connecting road segments. First, the network is constructed as a graph, where each intersection, such as People's Square or Century Avenue, is defined as a node, and the roads connecting these intersections, such as a section of Nanjing East Road, are defined as edges. The graph convolutional network model learns from this graph structure and understands how congestion at People's Square affects other road segments several kilometers away via the Yan'an Elevated Road. This impact is characterized by spatial dependency.

[0026] For example, for this section of Nanjing East Road, average speed and volume data were collected every five minutes over the past thirty days to form a long time series. This time series was fed into a gated recurrent unit model, which learned that this section of road experiences periodic peaks in travel time between 8:00 AM and 9:00 AM and 6:00 PM and 7:00 PM on weekdays, demonstrating a temporal variation pattern (i.e., a temporal periodic regularity feature). Ultimately, the model fuses the spatial dependency features extracted by the graph convolutional network with the temporal periodic regularity features extracted by the gated recurrent unit, thereby outputting precise travel time predictions for every five-minute window over the next two hours. For example, it predicts that it will take 300 seconds to travel through the section at 8:05 AM, but only 120 seconds at 10:10 AM.

[0027] S3, based on historical delivery data and time sensitivity, uses a Bayesian inference model to generate a discrete probability distribution of successful receipt for each order within the multiple time windows; Specifically, based on a large amount of historical delivery data (including destination location area, time period, and prior probability of successful signing for different destination types), the prior probability of successful signing for different areas, different time periods, and different destination types such as office buildings, residential areas, and shops is calculated.

[0028] For an order to be delivered, we construct a priori probability of successful receipt related to the order destination type and delivery time period. Based on the time sensitivity level of the order, we establish a likelihood function to measure the likelihood of the observed evidence (the time sensitivity of the order) given a certain hypothesis (for example, the order can be successfully received within a certain time period). The higher the time sensitivity level of the order, the higher the likelihood of receipt within the time window required by the order. We apply the Bayesian formula to integrate the prior probability of successful receipt and the likelihood function to obtain the posterior probability of successful receipt for the order in each future time window.

[0029] Specifically, suppose the system needs to evaluate an order destined for a residential community. Based on analysis of massive historical delivery data, the system determines that for this residential community's address type, the prior probability of successful delivery between 9:00 AM and 11:00 AM on weekdays is 70%, while the prior probability drops to 40% between 2:00 PM and 4:00 PM. This constitutes the prior probability of successful delivery for this order.

[0030] Specifically, if the order is for urgently needed medicine, its time sensitivity level is set to the highest level. The system will establish a likelihood function for this purpose. This function has an extremely high likelihood value, such as 0.95, within the time window specified by the customer for the order between 3:00 and 4:00 p.m., while the likelihood value is close to zero in other time windows. The Bayesian inference model then combines this high likelihood value with the lower 40% prior probability of successful receipt during the afternoon period to calculate an updated posterior probability of successful receipt. The final result may be that the probability of successful receipt in the window between 3:00 and 4:00 p.m. is greatly increased to 90%, while the probabilities in other time windows are correspondingly reduced, forming a probability distribution (i.e., generating a discrete probability distribution of successful receipt within the multiple time windows) to guide the delivery person to deliver at the optimal time. In other words, when the Bayesian formula combines the "lower 40% prior probability of successful receipt in the afternoon" with the "extremely high 0.95 likelihood value within the time window required by the order," the "probability of successful receipt in the window between 3:00 and 4:00 p.m. is significantly increased to 90%." This 90% is the posterior probability of successful receipt for the order in the window between 3:00 and 4:00 p.m.

[0031] S4, constructing a feasibility constraint set for vehicle mission execution based on the endurance, load limit, and health status index in the real-time vehicle status data; Specifically, the constraint set consists of a series of mathematical inequalities or logical constraints. For example, for any vehicle, the total weight of all assigned orders must not exceed the vehicle's rated load limit; the total mileage of its planned route must be less than the maximum range supported by the current remaining fuel or battery level; and the vehicle's health index, a comprehensive assessment of the condition of components such as the engine and tires, must be above a preset safety threshold, such as 80 points, to exclude the vehicle from the task assignment. These conditions collectively define the boundaries within which each vehicle can effectively execute a task.

[0032] In an optional embodiment, the feasibility constraint set for constructing vehicle task execution includes: setting an endurance constraint such that the total estimated energy consumption of the vehicle's planned route does not exceed the currently available energy; setting a load limit constraint such that the total weight and total volume of orders carried by the vehicle do not exceed the rated load and rated capacity of the vehicle; and setting a health status constraint such that the vehicle's health status index must be higher than a preset threshold.

[0033] For example, consider an electric delivery vehicle with 50 kWh of battery remaining, representing its available energy. Based on the total length of the planned route and the weight of the cargo on board, the system estimates the total energy consumption for this route to be 53 kWh. Because the estimated energy consumption of 53 kWh exceeds the available energy of 50 kWh, the route plan is deemed infeasible because it does not meet the range constraint.

[0034] Specifically, the rated load capacity of the vehicle is 500 kilograms and the rated volume is 4 cubic meters. The total weight of an order combination to be allocated is 480 kilograms, which meets the rated load capacity requirement, but its total volume is calculated to be 4.2 cubic meters, which exceeds the rated volume of the vehicle. Therefore, this order combination plan cannot be allocated to the vehicle because it does not meet the load limit constraint. In addition, each vehicle has a dynamically updated health status index ranging from zero to one hundred. The system sets a threshold of eighty. If a vehicle's health index drops to seventy-five due to severe tire wear or battery aging, any delivery plan involving this vehicle will be automatically rejected by the system for not meeting the health status constraint. The health status index can be pre-defined by technicians based on evaluation rules to obtain a score.

[0035] S5, taking the predicted travel time cost, the probability distribution of successful receipt, and the feasibility constraint set as input, applying a multi-objective optimization algorithm, with minimizing the predicted travel time cost of all vehicles and maximizing the total successful receipt probability of all orders as independent optimization goals, and solving to obtain a set of Pareto optimal delivery solutions containing multiple non-inferior solutions.

[0036] In an optional embodiment, the application of a multi-objective optimization algorithm to obtain a set of Pareto optimal distribution plan sets containing multiple non-inferior solutions includes: using a non-dominated sorting genetic algorithm (such as the NSGA-II algorithm) to encode potential distribution plans; setting a first objective function as the sum of the predicted travel time costs of all vehicle travel paths and minimizing it; setting a second objective function as the sum of the probabilities of successful receipt of all orders at their respective delivery time points and maximizing it; and generating a Pareto optimal solution set, i.e., a Pareto optimal distribution plan set, by iteratively performing selection, crossover, and mutation operations.

[0037] like Figure 2 Specifically, the algorithm aims to balance two conflicting objectives: total delivery time and overall sign-off success rate. A potential delivery plan, defining which driver delivers which packages and in what order, is encoded into a gene sequence. The algorithm first randomly generates one hundred such plans as an initial population. During the iteration process, the algorithm evaluates the total predicted travel time for each plan. For example, Plan A takes 420 minutes, while Plan B takes 500 minutes. The algorithm also calculates the overall sign-off success probability for each plan. For example, Plan A has an overall success rate of 85%, while Plan B has a success rate of 95%.

[0038] Specifically, the algorithm continuously evolves these solutions through repeated selection, crossover, and mutation operations. After hundreds of generations of evolution, the algorithm outputs a Pareto frontier solution set. This set may include Solution A, with a total time of 380 minutes and an 80% success rate; Solution B, with a total time of 430 minutes and a 92% success rate; and Solution C, with a total time of 510 minutes and a 98% success rate. No single solution is superior to another in all aspects. For example, Solution B has a higher success rate but is more time-consuming than Solution A, while Solution C takes less time but has a lower success rate. Together, these solutions constitute a set of high-quality alternatives representing different trade-offs.

[0039] S6, based on the preset risk preference or cost-effectiveness strategy, select the final delivery route from the Pareto optimal delivery solution set and send it to the vehicle terminal for execution.

[0040] like Figure 2Specifically, decision makers can set selection strategies based on their current business priorities. For example, during promotional events, to prioritize customer experience, a service-first strategy can be adopted, selecting the solution with the highest probability of successful sign-off from the Pareto optimal set. To control costs in daily operations, the solution with the lowest total travel time cost can be selected. Alternatively, the TOPSIS decision analysis method can be applied to quantitatively select a compromise solution that best balances these two objectives by calculating the relative proximity of each non-inferior solution to the ideal point (shortest time and highest probability) and the negative ideal point (longest time and lowest probability). After selecting a solution, the system packages each vehicle's specific driving route, customer visit sequence, and navigation instructions, and transmits them via the wireless network to the driver's mobile app or in-vehicle navigation device.

[0041] In an optional embodiment, the method of selecting a final delivery route from the Pareto optimal delivery solution set based on a preset risk preference or cost-effectiveness strategy includes: normalizing the total time cost (i.e., the sum of the predicted travel time costs) and the comprehensive sign-off success probability (i.e., the sum of the sign-off success probabilities) of each solution in the Pareto optimal delivery solution set; assigning corresponding weight coefficients to the normalized time cost and the comprehensive sign-off success probability based on the preset strategy; calculating a weighted comprehensive score for each solution based on the weight coefficients, and selecting the solution with the best comprehensive score as the final delivery route.

[0042] Specifically, suppose there are three solutions in the Pareto optimal set, with time costs of 300 minutes, 350 minutes, and 400 minutes, respectively, and corresponding comprehensive sign-off success probabilities of 85%, 92%, and 96%, respectively. The system first normalizes these values, mapping both the time cost and success probability to a range of zero to one for easier comparison.

[0043] Specifically, if the current business strategy prioritizes cost, such as for standard package delivery during off-peak hours, the operations manager can set a weight of 0.7 for time cost and a weight of 0.3 for the probability of successful signature. The system uses this weighted combination to calculate a weighted overall score for each option. In this case, the option with a cost of 300 minutes is likely to receive the highest score due to its significant time advantage. Conversely, if the delivery is for a batch of high-value, same-day expedited shipments, the strategy will prioritize service quality, with a weight of 0.2 for time cost and 0.8 for the probability of success. In this case, the option with a 96% probability of success, even if it takes longer, is likely to receive the highest overall score and be selected as the final delivery route.

[0044] like Figure 3As shown, the present invention also relates to a logistics distribution system based on the Internet of Things, including a processor and a memory, the memory storing a computer program, the processor can interact with the memory, can call the computer program (for example, through a bus), and then the processor executes the computer program. When the computer program is executed by the processor, the logistics distribution method based on the Internet of Things of the above embodiment is implemented.

[0045] Those skilled in the art will come up with many modifications, changes and substitutions without departing from the spirit and concept of the present invention. It should be understood that in practicing the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method of logistics distribution based on the Internet of Things, characterized in that: The following steps are involved: Obtain information about orders to be delivered, real-time status data of vehicles, road network topology of the target delivery area, and historical traffic flow data; Based on the road network topology and historical traffic flow data, a spatiotemporal prediction model integrating a graph convolutional network and a gated recurrent unit is used to calculate the predicted travel time cost of each road segment in the road network in multiple future time windows. Based on historical delivery data and time sensitivity, a Bayesian inference model is used to generate a discrete probability distribution of successful delivery for each order within the multiple time windows. Combining the vehicle's real-time status data with the vehicle's endurance, load limit, and health status index, a feasibility constraint set for the vehicle's mission execution is constructed; Taking the predicted travel time cost, the probability distribution of successful receipt, and the feasibility constraint set as input, a multi-objective optimization algorithm is applied, with minimizing the total predicted travel time cost of all vehicles and maximizing the total successful receipt probability of all orders as independent optimization objectives, to obtain a set of Pareto optimal delivery solutions containing multiple non-inferior solutions; According to the preset risk preference or cost-effectiveness strategy, the final delivery route is selected from the Pareto optimal delivery solution set and sent to the vehicle terminal for execution.

2. The method according to claim 1, characterized in that The spatiotemporal prediction model that integrates graph convolutional networks and gated recurrent units calculates the predicted travel time cost of each road segment in the road network in multiple future time windows, including: The road network topology is abstracted into a graph, with intersections as nodes and road sections as edges. The spatial dependency features of each road section are extracted using a graph convolutional network. The traffic flow data of each road section within a preset historical period is constructed into a time series and input into a gated recurrent unit to extract the temporal variation pattern. The spatial dependency features are integrated with the temporal variation pattern to output the predicted travel time cost for each time window in the future prediction domain with a preset duration as the unit.

3. The method according to claim 1, characterized in that The Bayesian inference model is used to generate a discrete probability distribution of successful receipt for each order within the multiple time windows, including: Based on historical delivery data, a priori probability of successful receipt related to the order destination type and delivery time period is constructed; according to the time sensitivity level of the order, a likelihood function is established, where the higher the time sensitivity level of the order, the higher the likelihood value of receipt within the time window required by the order; the Bayesian formula is applied to integrate the prior probability of successful receipt and the likelihood function to obtain the posterior probability of successful receipt of the order in each future time window.

4. The method according to claim 1, wherein The feasibility constraint set for constructing the vehicle mission execution includes: The cruising range constraint is set to ensure that the total estimated energy consumption of the vehicle's planned route does not exceed the currently available energy; the load limit constraint is set to ensure that the total weight and total volume of the orders carried by the vehicle do not exceed the vehicle's rated load and rated capacity; and the health status constraint is set to ensure that the vehicle's health status index must be higher than the preset threshold.

5. The method according to claim 1, wherein The multi-objective optimization algorithm is applied to obtain a set of Pareto optimal distribution solutions containing multiple non-inferior solutions, including: A multi-objective evolutionary algorithm based on non-dominated sorting is used to encode potential delivery plans. The first objective function is set as the sum of the predicted travel time costs of all vehicle routes and is minimized. The second objective function is set as the sum of the successful sign-in probabilities of all orders at their respective delivery time points and is maximized. The Pareto front solution set is generated by iteratively performing selection, crossover and mutation operations.

6. The method according to claim 1, characterized in that The selecting of the final delivery route from the Pareto optimal delivery solution set according to the preset risk preference or cost-effectiveness strategy includes: For each plan in the Pareto optimal delivery plan set, its total predicted travel time cost and total successful sign-off probability are normalized; according to the preset risk preference or cost-effectiveness strategy, corresponding weight coefficients are assigned to the normalized total predicted travel time cost and total successful sign-off probability; based on the weight coefficients, a weighted comprehensive score is calculated for each plan, and the plan with the best comprehensive score is selected as the final delivery route.

7. The method according to claim 1, characterized in that The order information includes order number, cargo weight, volume, delivery address and customer expected time.

8. The method according to claim 5, characterized in that The multi-objective evolutionary algorithm based on non-dominated sorting is the NSGA-II algorithm.

9. The method according to claim 6, characterized in that The risk appetite includes a service priority strategy.

10. A logistics distribution system based on the Internet of Things, characterized in that: The system comprises a processor and a memory, wherein the memory stores a computer program and the processor executes the computer program to implement the method for logistics distribution based on the Internet of Things as described in any one of claims 1 to 9.

Images