Double-body co-operation path optimization method based on improved NSGA-II

Abstract

The invention, which belongs to the logistics path calculation field, discloses an improved NSGA-II-based dual-subject co-transport path optimization method comprising the following steps of establishing a path planning problem model about dual-subject co-transport goods in the online car-hailing industry and the logistics industry based on profit and waiting time, and generating a path for each online hailed car, taking the path as a chromosome, iteratively updating the path in a crossing and variation mode, calculating the profit and waiting time of the updated path until the iteration frequency reaches a set suspender frequency threshold value, inputting the iterated path into the problem model, and outputting an optimal path. Partial express delivery can be delivered by the online car-hailing vehicle, the number of vehicles on the road is reduced to a certain extent, the road congestion is relieved, the emission of carbon dioxide is reduced, and for the online car-hailing vehicle, the no-load rate of the vehicle can be reduced, and extra profits can be obtained.

Description

Technical Field

[0001] The invention belongs to the field of logistics path calculation, and in particular relates to a dual-subject combined transport path optimization method based on improved NSGA-II. Background Art

[0002] With the booming development of e-commerce and the diversification of consumer demand, the logistics and transportation industry is facing unprecedented distribution pressure. At the same time, the high carbon emissions and high road congestion in the logistics industry have become urgent problems to be solved worldwide. The traditional logistics distribution model often relies on a large number of trucks to transport and distribute goods, which not only leads to a significant increase in carbon emissions, but also aggravates urban traffic congestion. In addition, some customers will choose to place orders through online car-hailing platforms to deliver their goods. However, in recent years, the saturation of the online car-hailing industry has continued to increase, and the phenomenon of online car-hailing vehicles running empty has become increasingly serious, which has further aggravated environmental problems and the uneven distribution of road resources. Against this background, an innovative distribution model has emerged, that is, online car-hailing platforms share part of the delivery tasks of express companies and realize collaborative delivery of goods.

[0003] Among the existing technical solutions, although some studies have explored the vehicle routing problem and its optimization methods, such as the green vehicle routing problem and the dynamic vehicle routing problem, these studies have made some progress in reducing carbon emissions and optimizing the number of vehicles. For example, some scholars have modeled and optimized carbon emissions by considering factors such as vehicle type, load, driving speed, and driving distance. However, most of these solutions are limited to vehicle routing optimization in a single industry and do not fully utilize the complementarity between different industries under the sharing economy. Summary of the invention

[0004] The purpose of the present invention is to provide a dual-subject combined transport path optimization method based on improved NSGA-II to solve the problems of significant increase in carbon emissions caused by the existing traditional logistics model, aggravated urban traffic congestion, and serious problem of online car-hailing vehicle empty driving.

[0005] In order to achieve the above object, the technical solution of the present invention is as follows: The present invention relates to a dual-agent combined transport path optimization method based on an improved NSGA-II, which comprises the following steps: S1. Based on the comprehensive consideration of profit and waiting time, a path planning model for the joint transportation of goods by the online car-hailing industry and the logistics industry is constructed; S2. The logistics distribution center, the customer's cargo pickup point, the customer's cargo drop-off point, and the express company's cargo are abstracted as nodes. For each online-hailing car, it is added to the path of the online-hailing car in sequence according to the distance between the customer's cargo pickup time window and the latest receipt time of the customer's cargo. Finally, a chromosome containing all the online-hailing car paths is formed and added to the chromosome set; S3. Set the iteration number threshold and set the initial iteration number to 0; S4. Set a crossover number threshold, select two chromosomes from the chromosome set to perform order-preserving crossover or double-vehicle crossover operations, form offspring chromosomes and add them to the chromosome set until the number of crossover operations reaches the set crossover number threshold; S5. traverse the offspring chromosomes and determine whether mutation is required. If necessary, perform mutation operation and add the offspring chromosomes after mutation operation to the chromosome set; S6. Calculate the profit and waiting time of the paths corresponding to all chromosomes, update the chromosomes in the chromosome set based on the profit and waiting time, add 1 to the number of iterations, and determine whether the number of iterations reaches the iteration threshold. If not, return to S4. If so, output the optimal path through the path planning problem model.

[0006] Preferably, S1 constructs a path planning problem model for the joint transportation of goods by the online car-hailing industry and the logistics industry, and the objective function of the problem model includes: (1), (2), in, F 1 is the total profit of all online ride-hailing vehicles, F 2 is the sum of the waiting time of all customers and the waiting time of all online car-hailing drivers, P Represents a collection of customer goods, p 1 for customer goods p The corresponding fees that customers need to pay, p 2 is the total delivery fee paid by the express company to all online car-hailing drivers, T Represents a collection of online ride-hailing vehicles. k Indicates the number of the online car-hailing vehicle. A is the carbon emission cost, B The fuel consumption cost of online ride-hailing vehicles, W Penalize the time window cost for online ride-hailing. p 0p For preferential customer goods p The fees that the customer should pay, ω For preferential treatment, μ For the cost of delivering a courier shipment, δ is the total number of delivered items, βis the unit fuel consumption cost, O For distribution centers, U A collection of pick-up points for customers’ goods. D A collection of drop-off points for customers’ goods. N It is a collection of distribution centers, customer cargo pickup points, customer cargo drop-off points, and express delivery company cargo receiving points. x ijk is a 0-1 variable, which means if the online car-hailing k Depend on i Point form to j point, then x ijk =1, otherwise, x ijk =0, d ij express i Click to j The distance between points, α represents the unit carbon emission cost, ε is the carbon emission rate per unit distance when the online car-hailing vehicle is unloaded and driving on a road section with a slope of 0. is the carbon emission rate per unit distance of an online car-hailing vehicle when fully loaded, load is the current load of the online car-hailing vehicle, is the maximum load capacity of online ride-hailing vehicles, GC is the road slope correction factor, M is the penalty base, OT p Goods for customers p If the boarding time window is exceeded, c For the number c Express company cargo, C For the collection of goods of express companies, OT c Cargo for express delivery companies c The time of the receiving time window, u For the number u Customer cargo pickup point, y pk is a 0-1 variable, which means if the customer's goods p By online car-hailing k Service y pk =1, otherwise y pk =0, AT uk For online ride-hailing k Arrive at the customer's cargo pickup point u time, ET p and LT pRespectively represent customer goods p The earliest and latest time of the boarding window.

[0007] Preferably, the problem model constructed in S1 further includes the following constraints: (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) in, z ck It is a 0-1 variable, which represents the express company's goods c By online car-hailing k Service z ck =1, otherwise z ck =0, AT Upk For online ride-hailing k Goods arriving at the customer p Pick-up point U time, AT Dpk For online ride-hailing k Goods arriving at the same customer p Drop-off point D time, AT Rck For online ride-hailing k Transport courier cargo c To the express company's customer cargo receiving point R time, q c Cargo for express delivery companies c The volume of Q The maximum volume that a ride-hailing vehicle can accommodate. Q 1 is the volume of each customer's cargo, S is a proper subset of the set N except the empty set, K For the collection of online ride-hailing vehicles, v is the average driving speed of online ride-hailing vehicles, IThe time threshold for the online car-hailing service to complete a round of tasks. RT c For goods c The latest receiving time.

[0008] Preferably, the specific steps of S2 forming the chromosome of each online car-hailing vehicle are: S2.1. For each online car-hailing vehicle, create a node that allows it to join; S2.2. Use the roulette method to select a node and estimate the arrival time of the online car-hailing vehicle; S2.3. Determine the node type. If the node is for express company cargo, proceed to S2.4-1. If the node is for customer cargo pickup point, proceed to S2.4-2. S2.4-1. Determine whether the time window requirement can be met. If so, add the node to the corresponding online car-hailing route, update the current time and the nodes allowed to join. If not, reset the current time and add a new route; S2.4-2. Determine whether the time window requirement can be met. If so, add the node and the corresponding drop-off point of the customer's goods to the corresponding online car-hailing route, update the current time and the nodes allowed to be added. If not, reset the current time and add a new route; The time window means that the arrival time of the online car-hailing vehicle must be between the time when the customer's goods arrive at the pick-up point and the latest acceptable pick-up time; S2.5. Determine whether the nodes allowed to join are empty. If not, return to S2.2. If it is empty, output the path to form a chromosome of an online car-hailing vehicle.

[0009] Preferably, in S2.4-2, after the node that can meet the time window requirement is added to the path of the corresponding online car-hailing vehicle, it is further determined whether the customer of the corresponding goods allows mid-way delivery. If allowed, the cargo node closest to the pick-up point is selected from the nodes allowed to be added to the path, and then the corresponding drop-off point of the customer's goods is added to the path of the corresponding online car-hailing vehicle; if not allowed, the drop-off point of the customer's goods is directly added to the path of the corresponding online car-hailing vehicle.

[0010] Preferably, the method of selecting the order-preserving crossing or the two-vehicle crossing operation in S4 is: respectively setting the probabilities of performing the order-preserving crossing and the two-vehicle crossing, and selecting the type of the crossing operation based on the probabilities of performing the order-preserving crossing and the two-vehicle crossing.

[0011] Preferably, the specific steps of the order-preserving crossover are: S4.1. Select at least one segment from the distribution center to the return distribution center from each of the two chromosomes as a crossover segment; S4.2. Keep all the fragments on the two chromosomes except the crossover fragment; S4.3. Add the nodes on the other chromosome that are not repeated in its own crossover fragment to the retained fragment in order to form two new daughter chromosomes.

[0012] Preferably, the specific steps of the two-vehicle crossing are: S4.4. Select a segment from each of the two chromosomes that starts from the distribution center and ends at the return distribution center and has the same number of nodes as a crossover segment; S4.5. Exchange the crossover segments on the two chromosomes and delete the nodes on the original chromosomes that are repeated with the crossover segments; S4.6. Supplement the deleted node on another chromosome to the vacant position of the deleted node on itself to form two new daughter chromosomes.

[0013] Preferably, the specific steps of the S5 mutation operation are: S5.1. Determine whether the receiving point of the express company's goods in the path corresponding to the chromosome is between the customer's goods pick-up point and the customer's goods unloading point. If so, perform mutation according to S5.2~S5.4; if not, perform mutation according to S5.5~S5.7; S5.2. Set the probability of the courier company's goods exchange mutation and the probability of the courier company's goods being moved to the customer's goods pick-up point; based on the probability, choose to perform the courier company's goods exchange mutation or the courier company's goods being moved to the customer's goods pick-up point mutation or not to perform the mutation. When the courier company's goods exchange mutation is selected, enter S5.3; when the courier company's goods are moved to the customer's goods pick-up point mutation, enter S5.4; S5.3. Exchange the nodes of the goods of the two express companies; S5.4. Move the node of one of the express delivery companies' goods to the node before a certain customer's goods are picked up; S5.5. Set the probability of moving the express company's goods to the customer's goods pickup point and drop-off point and the vehicle fusion mutation; based on the probability, choose to move the express company's goods to the customer's goods pickup point and drop-off point or the vehicle fusion mutation or no mutation. When the express company's goods are moved to the customer's goods pickup point and drop-off point, enter S5.6. When the vehicle fusion mutation is selected, enter S5.7. S5.6. Move the express company cargo node that is not between any customer cargo pickup and drop-off points to between a certain customer cargo pickup and drop-off point; S5.7. Delete the distribution point nodes between the two routes.

[0014] Preferably, after forming the chromosome set, S2 calculates the profits and waiting times of the paths corresponding to all chromosomes, and updates the chromosomes in the chromosome set based on the profits and waiting times. The specific steps of calculating the profits and waiting times of the paths corresponding to all chromosomes and updating the chromosomes in the chromosome set based on the profits and waiting times in S2 and S6 are: setting the scale of the chromosome set, calculating the profits and waiting times of the paths corresponding to the chromosomes, and performing non-dominated sorting on each path based on the profits and waiting times, that is, setting the path hierarchy. For two paths, when the profit corresponding to one path is greater than the profit corresponding to the other path and the waiting time is less than the waiting time corresponding to the other path, the hierarchy of the path is defined as lower than that of the other path, and the chromosomes corresponding to the paths are added to the chromosome set in order from low to high hierarchies until the number of chromosomes reaches the scale of the chromosome set.

[0015] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects: 1. The dual-subject joint transport path optimization method based on improved NSGA-II involved in the present invention constructs a path planning problem model for the dual-subject joint transport of goods in the online car-hailing industry and the logistics industry based on a comprehensive consideration of profit and waiting time, and then solves the problem model to form a feasible path plan. Some express deliveries can be delivered by online car-hailing companies, and express delivery companies only need to send fewer vehicles than before to meet their service requirements. This reduces the number of vehicles on the road to a certain extent, thereby alleviating road congestion and reducing carbon dioxide emissions.

[0016] 2. The dual-agent combined transport path optimization method based on improved NSGA-II involved in the present invention forms new path samples through continuous iterations of crossover operations and mutation operations when solving the path planning model, so as to find the optimal path with the highest profit and the least waiting time. The online car-hailing drivers can not only obtain more additional profits in almost the same time as in the traditional mode, but also reduce their own empty load rate, which has promoted and promoted the sharing economy to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG1 is a flow chart of a dual-agent combined transport path optimization method based on an improved NSGA-II; Figure 2 The flowchart of the formation process of the chromosome corresponding to the online car-hailing route; Figure 3 Schematic diagram of order-preserving crossover operation; Figure 4 It is a schematic diagram of a double vehicle crossing operation; Figure 5 A schematic diagram for exchanging the nodes of the express company's goods; Figure 6A schematic diagram of moving the node of one of the express companies' goods to the node before the customer's goods are picked up; Figure 7 A schematic diagram showing the movement of the express company's cargo node between the customer's cargo pickup and drop-off points; Figure 8 A schematic diagram of vehicle fusion; Fig. 9 It is a statistical diagram of the waiting time and profit of the optimal path obtained in the embodiment. DETAILED DESCRIPTION

[0018] In order to further understand the content of the present invention, the present invention is described in detail in conjunction with examples. The following examples are used to illustrate the present invention but are not used to limit the scope of the present invention.

[0019] See attached Figure 1 As shown, the present invention relates to a dual-subject combined cargo path optimization method based on an improved NSGA-II algorithm, comprising the following steps: S1. Based on the comprehensive consideration of profit and waiting time, a path planning problem model for the joint transportation of goods by the online car-hailing industry and the logistics industry is constructed. The objective function of the problem model includes: (1), (2), in, F 1 is the total profit of all online ride-hailing vehicles, F 2 is the sum of the waiting time of all customers and the waiting time of all online car-hailing drivers, P Represents a collection of customer goods, p 1 for customer goods p The corresponding fees that customers need to pay, p 2 is the total delivery fee paid by the express company to all online car-hailing drivers, T Represents a collection of online ride-hailing vehicles. k Indicates the number of the online car-hailing vehicle. A is the carbon emission cost, B The fuel consumption cost of online ride-hailing vehicles, W Penalize the time window cost for online ride-hailing. p 0p For preferential customer goods p The fees that the customer should pay, ω For preferential treatment, μ For the cost of delivering a courier shipment, δ is the total number of delivered items, β is the unit fuel consumption cost, O For distribution centers, U A collection of pick-up points for customers’ goods. DA collection of drop-off points for customers’ goods. N It is a collection of distribution centers, customer cargo pickup points, customer cargo drop-off points, and express delivery company cargo receiving points. x ijk is a 0-1 variable, which means if the online car-hailing k Depend on i Point form to j point, then x ijk =1, otherwise, x ijk =0, d ij express i Click to j The distance between points, α represents the unit carbon emission cost, ε is the carbon emission rate per unit distance when the online car-hailing vehicle is unloaded and driving on a road section with a slope of 0. is the carbon emission rate per unit distance of an online car-hailing vehicle when fully loaded, load is the current load of the online car-hailing vehicle, is the maximum load capacity of online ride-hailing vehicles, GC is the road slope correction factor, M is the penalty base, OT p Goods for customers p If the boarding time window is exceeded, c For the number c Express company cargo, C For the collection of goods of express companies, OT c Cargo for express delivery companies c The time of the receiving time window, u For the number u Customer cargo pickup point, y pk is a 0-1 variable, which means if the customer's goods p By online car-hailing k Service y pk =1, otherwise y pk =0, AT uk For online ride-hailing k Arrive at the customer's cargo pickup point u time, ET p and LT p Respectively represent customer goods p The earliest and latest time of the boarding window.

[0020] The above problem model also includes the following constraints: (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) in, z ck It is a 0-1 variable, which represents the express company's goods c By online car-hailing k Service z ck =1, otherwise z ck =0, U p and D p Respectively represent customer goods p Pick-up and drop-off points, AT Upk For online ride-hailing k Goods arriving at the customer p Pick-up point U time, AT Dpk For online ride-hailing k Goods arriving at the same customer p Drop-off point D time, R c Express company cargo c The receiving point, AT Rck For online ride-hailing k Transport courier cargo c To the express company's customer cargo receiving point R time, q c Cargo for express delivery companies c The volume of Q The maximum volume that a ride-hailing vehicle can accommodate. Q 1 is the volume of each customer's cargo, S is a proper subset of the set N except the empty set, KFor the collection of online ride-hailing vehicles, v is the average driving speed of online ride-hailing vehicles, I The time threshold for the online car-hailing service to complete a round of tasks. RT c For goods c The latest receiving time; Formula (3) and (4) indicate that all nodes except the distribution center can only be passed once; Formula (5) indicates that any customer's goods must be delivered from the pick-up point to the drop-off point by the same vehicle; Formula (6) indicates that any express company's goods can only be delivered by one vehicle, and the vehicle responsible for delivering a certain goods is determined at the beginning and cannot be transferred in the middle; Formula (7) indicates that the time when the online car-hailing vehicle arrives at the customer's goods pick-up point is before the time when it arrives at the same customer's goods drop-off point; Formula (8) indicates that the time when the online car-hailing vehicle arrives at the customer's goods pick-up point must be before the latest pick-up time of the customer's goods; Formula (9) indicates that the online car-hailing vehicle must deliver the goods before the latest receiving time of the express company's goods; Formula (10) is the volume constraint of the goods taken out by each online car-hailing vehicle from the distribution center; The purpose of Formula (11) is to eliminate sub-loops; Formula (12) indicates that the time for each online car-hailing vehicle to complete a round of tasks cannot exceed I .

[0021] S2. The logistics distribution center, customer cargo pick-up point, customer cargo unloading point and cargo are abstracted as nodes. For each online-hailing car, the nodes to be added to the online-hailing car are selected according to the distance between the customer cargo pick-up time window and the latest receipt time of the customer cargo, forming the chromosome of each online-hailing car and adding it to the chromosome set.

[0022] The present invention adopts the natural number coding method to assign specific numbers to the logistics distribution center, each customer cargo loading point, customer cargo unloading point and express company cargo unloading point. n Indicates the total number of customer goods, m Indicates the total number of goods of the express company, the number "0" is used to represent the distribution center, and the number "1- n " indicates the pick-up point for the customer's goods, and the number " n +1-2 n " indicates the drop-off point for the corresponding customer's goods, and the number "2 n +1-2 n + m " indicates the express company's goods, and the nodes are numbered. n = m=30 as an example, for a chromosome of an online car-hailing vehicle of 0-4-78-34-0-80-2-32-0, it means: an online car-hailing vehicle takes out cargo No. 78 from the distribution center, first arrives at the pick-up point for customer No. 4's cargo, delivers cargo No. 78 on the way to the drop-off point 34 for the customer's cargo, then departs from the drop-off point 34 and returns to the distribution center; another vehicle takes out cargo No. 80 from the distribution center and starts delivering it, and after delivering the cargo, it goes to the pick-up point for customer No. 2's cargo, delivers it from the pick-up point to the drop-off point 32, and finally returns to the distribution center.

[0023] See attached Figure 2 As shown, the specific steps for forming the chromosome of each online car-hailing vehicle are: S2.1. For each online car-hailing vehicle, create a node that allows it to join; S2.2. Use roulette (existing selection algorithm) to select a node and estimate the arrival time of the online car-hailing vehicle; S2.3. Determine the node type. If the node is cargo, proceed to S2.4-1. If the node is a customer cargo pickup point, proceed to S2.4-2. S2.4-1. Determine whether the time window requirement can be met. If so, add the node to the corresponding online car-hailing route, update the current time and the nodes allowed to join. If not, reset the current time and add a new route; S2.4-2. Determine whether the time window requirement can be met. If not, reset the current time and add a new path. If it can be met, add the node and the corresponding drop-off point of the customer's goods to the path of the corresponding online car-hailing vehicle. Considering the different needs of each user, the customer can decide whether to allow the online car-hailing vehicle to deliver other goods midway. If allowed, select the cargo node closest to the pick-up point from the nodes allowed to be added to the path, and then add the corresponding drop-off point of the customer's goods to the path of the corresponding online car-hailing vehicle; if not allowed, directly add the drop-off point of the customer's goods to the path of the corresponding online car-hailing vehicle; update the current time and nodes allowed to be added; The time window means that the arrival time of the online car-hailing vehicle must be between the time when the customer's goods arrive at the pick-up point and the latest acceptable pick-up time; S2.5. Determine whether the nodes allowed to join are empty. If not, return to S2.2. If it is empty, output the path to form a chromosome of an online car-hailing vehicle.

[0024] After repeating S2.1-S2.5, all chromosomes are generated and a chromosome set is established, that is, the initial solution is constructed.

[0025] According to formula (1) and formula (2), the profits and waiting times of the paths corresponding to all chromosomes are calculated, and the chromosomes in the chromosome set are updated based on the profits and waiting times. The specific steps are as follows: set the size of the chromosome set, perform non-dominated sorting on each path based on the profits and waiting times, that is, set the path hierarchy. For two paths, when the profit corresponding to one path is greater than the profit corresponding to the other path and the waiting time is less than the waiting time corresponding to the other path, the hierarchy of the path is defined as lower than the hierarchy of the other path. When the profit and waiting time of one chromosome are both greater than the other chromosome, or the profit and waiting time of one chromosome are both less than the other chromosome, the two chromosomes are defined as being at the same hierarchy. The chromosomes corresponding to the paths are added to the chromosome set in order from low to high hierarchies until the number of chromosomes reaches the size of the chromosome set. In this way, only the chromosomes corresponding to the paths with relatively optimized profits and waiting times are retained.

[0026] S3. Set the iteration number threshold and set the initial iteration number to 0; S4. Set a crossover number threshold, and select two chromosomes from the chromosome set by a binary tournament selection method (this selection method belongs to the conventional technical means in this field) to perform order-preserving crossover or double-vehicle crossover operation. The two crossover operation selection methods are: respectively set the probability of performing order-preserving crossover and double-vehicle crossover, for example, set a probability of 60% for performing order-preserving crossover and a probability of 40% for performing double-vehicle crossover, and select the type of crossover operation based on the probability of performing order-preserving crossover and double-vehicle crossover.

[0027] See attached Figure 3 As shown, the specific steps of the order-preserving crossover are: S4.1. The two selected chromosomes are used as parent 1 and parent 2, and at least one segment from the distribution center to the return distribution center is selected from each of the two chromosomes as a crossover segment; S4.2. Keep all the fragments on the two chromosomes except the crossover fragment; S4.3. Add the nodes on the other chromosome that are not repeated in its own crossover fragment to the retained fragment in order to form two new offspring chromosomes, namely offspring 1 and offspring 2.

[0028] See attached Figure 4 As shown, the specific steps of the double vehicle crossing are: S4.4. The two selected chromosomes are used as parent 1 and parent 2, and a segment with the same number of nodes from the distribution center to the return distribution center is selected from each of the two chromosomes as a crossover segment; S4.5. Exchange the crossover segments on the two chromosomes and delete the nodes on the original chromosomes that are repeated with the crossover segments; S4.6. Supplement the deleted node on another chromosome to the vacant position of the deleted node on itself to form two new offspring chromosomes, namely offspring 1 and offspring 2.

[0029] The offspring chromosomes formed after the crossover operation are added to the chromosome set, and the step S4 is repeated until the number of crossover operations reaches the set crossover threshold. The purpose of the crossover operation is to enrich the diversity of the population and expand the search space.

[0030] S5. Traverse the offspring chromosomes and determine whether mutation is required. If necessary, perform mutation operation and add the offspring chromosomes after mutation operation to the chromosome set. The specific steps of mutation operation are: S5.1. All the express delivery companies' goods on the chromosome can be divided into two categories: (1) Determine whether the express delivery company's goods receiving point in the path corresponding to the chromosome is between the customer's goods pick-up point and the customer's goods drop-off point; (2) Determine whether the express delivery company's goods receiving point in the path corresponding to the chromosome is not between the customer's goods pick-up point and the customer's goods drop-off point. Therefore, first determine whether the express delivery company's goods receiving point in the path corresponding to the chromosome is between the customer's goods pick-up point and the customer's goods drop-off point. If so, perform mutation according to S5.2~S5.4. If not, that is, the express delivery company's goods receiving point is not between the customer's goods pick-up point and the customer's goods drop-off point, perform mutation according to S5.5~S5.7. S5.2. Set the probability of the courier company's goods exchange mutation and the probability of the courier company's goods being moved to the customer's goods pick-up point. For example, the probability of the courier company's goods exchange mutation is 15%, and the probability of the courier company's goods being moved to the customer's goods pick-up point is 5%. When the courier company's goods exchange mutation is selected, enter S5.3. When the courier company's goods are moved to the customer's goods pick-up point, enter S5.4. S5.3. Exchange the nodes of the goods of the two express companies, such as Figure 5 As shown; S5.4. Move the node of one of the express delivery companies' goods to the node before a customer's goods loading node, such as Figure 6 As shown; S5.5. Set the probability of moving the express company's goods to the customer's goods pickup point and drop-off point and the vehicle fusion mutation; based on the probability, choose to move the express company's goods to the customer's goods pickup point and drop-off point or the vehicle fusion mutation or no mutation. When the express company's goods are moved to the customer's goods pickup point and drop-off point, enter S5.6. When the vehicle fusion mutation is selected, enter S5.7. S5.6. Move the express company cargo node that is not between any customer cargo pickup and drop-off points to a certain customer cargo pickup and drop-off point, such as Figure 7 As shown; S5.7. Delete the distribution point nodes between the two paths, such as Figure 8 shown.

[0031] S6. Calculate the profits and waiting times of the paths corresponding to all chromosomes, and perform non-dominated sorting again, that is, update the chromosomes in the chromosome set based on the profits and waiting times, set the path hierarchy, and for two paths, when the profit corresponding to one path is greater than the profit corresponding to the other path and the waiting time is less than the waiting time corresponding to the other path, define the hierarchy of the path as lower than the hierarchy of the other path, and add the chromosomes corresponding to the paths to the chromosome set in order from low to high until the number of chromosomes reaches the size of the chromosome set, add 1 to the number of iterations, and determine whether the number of iterations reaches the iteration threshold. If not, return to S4. If reached, output the optimal path through the path planning problem model.

[0032] In order to prove that the method of the present invention has a significant improvement in the performance of the initial solution of the problem model, this embodiment statistically analyzes the waiting time and profit of the optimal path obtained by solving the path set before step S3 and the waiting time and profit of the optimal path obtained after step S6, as shown in FIG. Fig. 9 As shown, the horizontal axis is the waiting time, the unit is "minutes", and the vertical axis is the profit, the unit is "yuan". It can be seen that both the total waiting time of customers and drivers and the profit of online car-hailing have been significantly improved.

[0033] The present invention is described in detail above in conjunction with the embodiments, but the contents described are only preferred embodiments of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of application of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A dual-agent combined transport path optimization method based on improved NSGA-II, characterized by: It includes the following steps: S1. Based on the comprehensive consideration of profit and waiting time, a path planning model for the joint transportation of goods by the online car-hailing industry and the logistics industry is constructed; S2. The logistics distribution center, the customer's cargo pickup point, the customer's cargo drop-off point, and the express company's cargo are abstracted as nodes. For each online-hailing car, it is added to the path of the online-hailing car in sequence according to the distance between the customer's cargo pickup time window and the latest receipt time of the customer's cargo. Finally, a chromosome containing all the online-hailing car paths is formed and added to the chromosome set; S3. Set the iteration number threshold and set the initial iteration number to 0; S4. Set a crossover number threshold, select two chromosomes from the chromosome set to perform order-preserving crossover or double-vehicle crossover operations, form offspring chromosomes and add them to the chromosome set until the number of crossover operations reaches the set crossover number threshold; S5. traverse the offspring chromosomes and determine whether mutation is required. If necessary, perform mutation operation and add the offspring chromosomes after mutation operation to the chromosome set; S6. Calculate the profit and waiting time of the paths corresponding to all chromosomes, update the chromosomes in the chromosome set based on the profit and waiting time, add 1 to the number of iterations, and determine whether the number of iterations reaches the iteration threshold. If not, return to S4. If so, output the optimal path through the path planning problem model.

2. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: S1 constructs a path planning problem model for the joint transportation of goods by the online car-hailing industry and the logistics industry. The objective function of the problem model includes: （1）， （2）， in, F 1 is the total profit of all online ride-hailing vehicles, F 2 is the sum of the waiting time of all customers and the waiting time of all online car-hailing drivers, P Represents a collection of customer goods, p 1 for customer goods p The corresponding fees that customers need to pay, p 2 is the total delivery fee paid by the express company to all online car-hailing drivers, T Represents a collection of online ride-hailing vehicles. k Indicates the number of the online car-hailing vehicle. A is the carbon emission cost, B is the fuel consumption cost of online ride-hailing vehicles, W Penalize the time window cost for online ride-hailing. p 0p For preferential customer goods p The fees that the customer should pay, ω For preferential treatment, μ For the cost of delivering a courier shipment, δ is the total number of deliveries, β is the unit fuel consumption cost, O For distribution centers, U A collection of pick-up points for customers’ goods. D A collection of drop-off points for customers’ goods. N It is a collection of distribution centers, customer cargo pickup points, customer cargo drop-off points, and express delivery company cargo receiving points. x ijk is a 0-1 variable, which means if the online car-hailing k Depend on i Point form to j point, then x ijk =1, otherwise, x ijk =0, d ij express i Click to j The distance between points, α represents the unit carbon emission cost, ε is the carbon emission rate per unit distance when the online car-hailing vehicle is unloaded and driving on a road section with a slope of 0. is the carbon emission rate per unit distance of an online car-hailing vehicle when fully loaded, load is the current load of the online car-hailing vehicle, is the maximum load capacity of online ride-hailing vehicles, GC is the road slope correction factor, M is the penalty base, OT p Goods for customers p If the boarding time window is exceeded, c For the number c Express company cargo, C For the collection of express company goods, OT c Cargo for express delivery companies c The time of the receiving time window, u For the number u Customer cargo pickup point, y pk is a 0-1 variable, which means if the customer's goods p By online car-hailing k Service y pk =1, otherwise y pk =0, AT uk For online ride-hailing k Arrive at the customer's cargo pickup point u time, ET p and LT p Respectively represent customer goods p The earliest and latest time of the boarding window.

3. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 2 is characterized by: The problem model constructed in S1 also includes the following constraints: (3)， (4)， (5)， (6)， (7)， (8)， (9)， (10)， (11)， (12)， in, z ck It is a 0-1 variable, which represents the express company's goods c By online car-hailing k Service z ck =1, otherwise z ck =0, AT Upk For online ride-hailing k Goods arriving at the customer p Pick-up point U time, AT Dpk For online ride-hailing k Goods arriving at the same customer p Drop-off point D time, AT Rck For online ride-hailing k Transport courier cargo c To the courier company's customer cargo receiving point R time, q c Cargo for express delivery companies c The volume of Q The maximum volume that an online car-hailing vehicle can accommodate. Q 1 is the volume of each customer's cargo, S is a proper subset of the set N excluding the empty set, K For the collection of online ride-hailing vehicles, v is the average driving speed of online ride-hailing vehicles, I The time threshold for the online car-hailing service to complete a round of tasks. RT c For goods c The latest receiving time.

4. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: The specific steps of S2 forming the chromosome of each online car-hailing vehicle are: S2.

1. For each online car-hailing vehicle, create a node that allows it to join; S2.

2. Use the roulette method to select a node and estimate the arrival time of the online car-hailing vehicle; S2.

3. Determine the node type. If the node is for express company cargo, proceed to S2.4-1. If the node is for customer cargo pickup point, proceed to S2.4-2. S2.4-1. Determine whether the time window requirement can be met. If so, add the node to the corresponding online car-hailing route, update the current time and the nodes allowed to join. If not, reset the current time and add a new route; S2.4-2. Determine whether the time window requirement can be met. If so, add the node and the corresponding drop-off point of the customer's goods to the corresponding online car-hailing route, update the current time and the nodes allowed to be added. If not, reset the current time and add a new route; The time window means that the arrival time of the online car-hailing vehicle must be between the time when the customer's goods arrive at the pick-up point and the latest acceptable pick-up time; S2.

5. Determine whether the nodes allowed to join are empty. If not, return to S2.

2. If it is empty, output the path to form a chromosome of an online car-hailing vehicle.

5. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 4 is characterized by: In S2.4-2, after the node that can meet the time window requirement is added to the path of the corresponding online car-hailing vehicle, it is further determined whether the customer of the corresponding goods allows mid-way delivery. If allowed, the cargo node closest to the pick-up point is selected from the nodes allowed to be added to the path, and then the corresponding drop-off point of the customer's goods is added to the path of the corresponding online car-hailing vehicle; if not allowed, the drop-off point of the customer's goods is directly added to the path of the corresponding online car-hailing vehicle.

6. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: The method of selecting the order-preserving crossing or the two-vehicle crossing operation in S4 is: respectively setting the probabilities of performing the order-preserving crossing and the two-vehicle crossing, and selecting the type of the crossing operation based on the probabilities of performing the order-preserving crossing and the two-vehicle crossing.

7. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: The specific steps of the order-preserving crossover are: S4.

1. Select at least one segment from the distribution center to the return distribution center from each of the two chromosomes as a crossover segment; S4.

2. Keep all the fragments on the two chromosomes except the crossover fragment; S4.

3. Add the nodes on the other chromosome that are not repeated in its own crossover fragment to the retained fragment in order to form two new daughter chromosomes.

8. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: The specific steps of the double vehicle crossing are: S4.

4. Select a segment from each of the two chromosomes that starts from the distribution center and ends at the return distribution center and has the same number of nodes as a crossover segment; S4.

5. Exchange the crossover segments on the two chromosomes and delete the nodes on the original chromosomes that are repeated with the crossover segments; S4.

6. Supplement the deleted node on another chromosome to the vacant position of the deleted node on itself to form two new daughter chromosomes.

9. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: The specific steps of the S5 mutation operation are: S5.

1. Determine whether the receiving point of the express company's goods in the path corresponding to the chromosome is between the customer's goods pick-up point and the customer's goods unloading point. If so, mutate according to S5.2~S5.

4. If not, mutate according to S5.5~S5.

7. S5.

2. Set the probability of the courier company's goods exchange mutation and the probability of the courier company's goods being moved to the customer's goods pick-up point; based on the probability, choose to perform the courier company's goods exchange mutation or the courier company's goods being moved to the customer's goods pick-up point mutation or not to perform the mutation. When the courier company's goods exchange mutation is selected, enter S5.3; when the courier company's goods are moved to the customer's goods pick-up point mutation, enter S5.4; S5.

3. Exchange the nodes of the goods of the two express companies; S5.

4. Move the node of one of the express delivery companies' goods to the node before a certain customer's goods are picked up; S5.

5. Set the probability of variation between the express company's goods moving to the customer's goods pick-up and drop-off points and the variation of vehicle fusion; Based on the probability, the variation of moving the express company's goods between the customer's goods pick-up point and the drop-off point or the variation of vehicle fusion or no variation is selected. When the variation of moving the express company's goods between the customer's goods pick-up point and the drop-off point is selected, enter S5.6, and when the variation of vehicle fusion is selected, enter S5.7; S5.

6. Move the express company cargo node that is not between any customer cargo pickup and drop-off points to between a certain customer cargo pickup and drop-off point; S5.

7. Delete the delivery point nodes between the two routes.

10. The dual-agent combined transport path optimization method based on improved NSGA-II according to claim 1 is characterized by: After the chromosome set is formed in S2, the profits and waiting times of the paths corresponding to all chromosomes are calculated, and the chromosomes in the chromosome set are updated based on the profits and waiting times. The specific steps of calculating the profits and waiting times of the paths corresponding to all chromosomes and updating the chromosomes in the chromosome set based on the profits and waiting times in S2 and S6 are as follows: setting the scale of the chromosome set, calculating the profits and waiting times of the paths corresponding to the chromosomes, and performing non-dominated sorting on each path based on the profits and waiting times, that is, setting the path hierarchy. For two paths, when the profit corresponding to one path is greater than the profit corresponding to the other path and the waiting time is less than the waiting time corresponding to the other path, the hierarchy of the path is defined to be lower than that of the other path, and the chromosomes corresponding to the paths are added to the chromosome set in order from low to high hierarchies until the number of chromosomes reaches the scale of the chromosome set.

Images