Multi-string-point path planning method and system based on urban distribution scene

Abstract

The invention discloses a multi-string-point path planning method and system based on an urban distribution scene, and the method comprises the steps: carrying out the modeling and integration of a traditional path planning model and a three-dimensional loading model, and building a multi-string-point path planning and 3D loading integrated model which takes the total cost, a time penalty function and the total load factor of a vehicle as multiple targets. On the basis of constraint conditions, the influence of the cargo loading sequence and the vehicle cargo taking path on key operation factors such as the vehicle loading rate and the transportation cost is considered at the same time, a VRP and 3D-CLP integrated optimization result is obtained, the balanced optimal decision of the vehicle path and the cargo loading mode is achieved, the safety and the loading rate of a loading and transportation system are improved, and the loading and transportation efficiency is improved. And the damage rate of goods is obviously reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of logistics, and specifically relates to a multi-stop path planning method and system based on the urban distribution scenario. Background Art

[0002] With the rapid development of e-commerce and retail, the society's demand for efficient and flexible distribution services for urban distribution logistics in the 2B segment (a logistics distribution service model within the city mainly targeting enterprise customers) is increasing day by day. In the 2B segment of urban distribution, distribution scenarios dominated by commodities such as home appliances, furniture, and fast-moving consumer goods face severe challenges due to the complex characteristics of multi-point pick-up and multi-point delivery. This not only requires urban distribution enterprises to have a high degree of supply chain coordination ability, accurately plan pick-up and delivery routes, but also needs to improve operational efficiency and flexibility.

[0003] Different from single-point distribution and C-end timely distribution, the key to B-end urban distribution lies in scheduling, mainly starting from two aspects: vehicle loading and path planning. In the urban distribution scenario, there is a close association and influence between the order of loading goods on the vehicle and the selection of the pick-up points and delivery points of the vehicle. If the vehicle routing problem (abbreviated as VRP) and the three-dimensional container loading problem (abbreviated as 3D-CLP) in this scenario are split into two independent problems for solution, no matter which problem is solved first, it may lead to only a local optimal solution for the second problem to be solved. Summary of the Invention

[0004] The present invention proposes a multi-stop path planning method and system based on the urban distribution scenario, integrates the traditional path planning model and the 3D loading model by finding the association, establishes a multi-stop path planning and three-dimensional loading integrated model with the total cost, time penalty function, and total vehicle full-load rate as multiple objectives, focuses on considering real factors, adapts to the pain points of actual logistics urban distribution problems, and finally realizes the integrated optimization of VRP and 3D-CLP.

[0005] The present invention is implemented by the following technical solutions:

[0006] A multi-stop path planning method based on the urban distribution scenario is proposed, including:

[0007] Construct a total cost objective function for the vehicle delivery process with the minimum of the fixed cost and variable cost of the vehicle as the goal:

[0008]

[0009] D k is the fixed cost of calling vehicle k, xijk Indicate whether vehicle k drives from node i to node j. If so, it is 1; otherwise, it is 0; C k is the fuel cost consumed by vehicle k per kilometer on average; d ij Indicates the driving distance from node i to node j;

[0010] Construct a time penalty function with the goal of minimizing the cost of violating the order timeliness:

[0011] Among them,

[0012] Among them, Indicates the expected arrival time of the customer for the order picked up at node i; T i 2 Indicates the latest arrival time of the order picked up at node i; p represents the penalty cost coefficient for the goods arriving after their soft time window;

[0013] Construct a total vehicle full load rate function with the goal of maximizing the volume full load rate and the load full load rate:

[0014]

[0015] ; Among them, A ink Is used to indicate whether the goods picked up at node i are loaded in the carriage when vehicle k leaves node n. If so, it is 1; otherwise, it is 0; Indicates the weight of the m-th piece of goods picked up at node i; x njk Indicates whether vehicle k drives from node n to node j; W k Indicates the rated load capacity of vehicle K; V k Indicates the rated volume of vehicle K;

[0016] Solve the total cost objective function, the time penalty function, and the total vehicle full load rate function based on the NSGA non-dominated genetic algorithm to obtain the optimal distribution path for multi-point pickup and multi-point delivery.

[0017] In some embodiments of the present invention, the constraint conditions of the vehicle routing problem include:

[0018] P∩D=0;

[0019]

[0020] 0≤x ijk ≤ME ijk ,i∈I∪P∪D,j∈P∪D,k∈K;

[0021] a ik <a (N+i)k ,i∈P;

[0022]

[0023]

[0024] Among them, P = {1, 2,... N} is the set of the positions of all pick-up points in the distribution network; D = {N + 1, N + 2,... 2N} is the set of the positions of all delivery points in the distribution network; y ik indicates whether the goods at the i-th pick-up point are delivered by vehicle k. If so, it is 1; otherwise, it is 0; E ijk indicates whether vehicle k can pass through the section from node i to node j. If it can, it is 1; otherwise, it is 0; a ik is the time when vehicle k arrives at node i; r ik indicates the time when vehicle k leaves node i; y ik indicates whether the goods at the i-th pick-up point are delivered by vehicle k. If so, it is 1; otherwise, it is 0.

[0025] In some embodiments of the present invention, the three-dimensional loading constraint conditions include:

[0026]

[0027] For all i p ≠j q , it is necessary to satisfy one of each of the following three groups of expressions:

[0028]

[0029] For all it is necessary to satisfy the following two expressions:

[0030]

[0031] If the goods i p are unloaded before the goods j q , it is necessary to satisfy the following group of expressions:

[0032]

[0033] Among them, represents the coordinates of the left rear lower vertex of the m-th piece of goods taken by vehicle k from node i; is the upper right front vertex coordinate of the goods i m on the vehicle k; is the length, width, and height of the goods i m ; expresses the state that the goods i p and the goods j q are both in the carriage of vehicle k. If the goods i p and the goods j q are both in the carriage of vehicle k, it is 1; otherwise, it is 0.

[0034] In some embodiments of the present invention, before solving the optimal delivery route using the NSGA non-dominated sorting genetic algorithm, the method further includes:

[0035] Clustering according to the geographical locations of the pick-up points and delivery points of the orders using the DBSCAN clustering method, so that the orders in the same clustering cluster are delivered by one vehicle.

[0036] In some embodiments of the present invention, the method further includes:

[0037] Design a loading engine module to simulate the three-dimensional packing of goods, record the available space, the storage coordinates and order of the goods, and the loading process satisfies various constraint conditions in the model;

[0038] Design a feasibility inspection module to check the order of the pick-up points and delivery points of the route, and check whether the order is completed in combination with the loading engine, and regard the routes that all meet the requirements as feasible routes.

[0039] A multi-stop route planning system based on the urban distribution scenario is proposed, including:

[0040] A total cost construction unit for constructing a total cost objective function for the vehicle delivery process with the goal of minimizing the fixed cost and variable cost of the vehicle:

[0041]

[0042] Among them, D k is the fixed cost of calling vehicle k, and x ijk represents whether vehicle k drives from node i to node j; if so

[0043] is 1, otherwise it is 0; C k is the fuel cost consumed by vehicle k per kilometer on average; d ij represents the driving distance from node i to node j;

[0044] A time penalty function construction unit for constructing a time penalty function with the goal of minimizing the cost of violating the order timeliness:

[0045] Among them,

[0046] Among them, T i 1 represents the expected arrival time of the customer of the order picking up goods from node i; T i 2 represents the latest arrival time of the order picking up goods from node i; p represents the penalty cost coefficient for the goods arriving later than their soft time window;

[0047] A vehicle total full load rate function construction unit is used to construct a vehicle total full load rate function with the highest volume full load rate and load full load rate:

[0048] where A ink is used to represent whether the goods taken out from node i are loaded in the carriage when vehicle k leaves node n. If so, it is 1; otherwise, it is 0; w im represents the weight of the m-th piece of goods taken from node i; x njk represents whether vehicle k drives from node n to node j; W k represents the rated load capacity of vehicle K; V k represents the rated volume of vehicle K;

[0049] An optimal delivery route solving unit is used to solve the total cost objective function, time penalty function, and vehicle total full load rate function based on the NSGA non-dominated genetic algorithm to obtain the optimal delivery route for multi-point pick-up and multi-point delivery.

[0050] In some embodiments of the present invention, the constraint conditions of the vehicle routing problem include:

[0051] P ∩ D = 0;

[0052]

[0053] 0 ≤ x ijk ≤ ME ijk , i ∈ I ∪ P ∪ D, j ∈ P ∪ D, k ∈ K;

[0054] a ik < a (N+i)k , i ∈ P;

[0055]

[0056] where P = {1, 2,... N} is the set of positions of all pick-up points in the distribution network; D = {N + 1, N + 2,... 2N} is the set of positions of all delivery points in the distribution network; y ik represents whether the goods at the i-th pick-up point are delivered by vehicle k. If so, it is 1; otherwise, it is 0; E ijk represents whether vehicle k can pass through the section from node i to node j. If it can, it is 1; otherwise, it is 0; a ik is the time when vehicle k arrives at node i; r ik represents the time when vehicle k leaves node i; y ik represents whether the goods at the i-th pick-up point are delivered by vehicle k. If so, it is 1; otherwise, it is 0.

[0057] In some embodiments of the present invention, the three-dimensional loading constraint conditions include:

[0058]

[0059] For all i p ≠j q , one of each of the following three groups of expressions must be satisfied:

[0060]

[0061] For all , the following two expressions must be satisfied:

[0062]

[0063]

[0064] If cargo i p is unloaded before cargo j q , the following group of expressions must be satisfied:

[0065]

[0066] Wherein, represents the left rear lower vertex coordinates of the m-th cargo taken by vehicle k from node i; is the right front upper vertex coordinates of cargo i m on vehicle k; is the length, width and height of cargo i m ; expresses the state of cargo i p and cargo j q being in the carriage of vehicle k at the same time. If cargo i p and cargo j q are in the carriage of vehicle k at the same time, it is 1, otherwise it is 0.

[0067] In some embodiments of the present invention, before the optimal delivery route solving unit solves the optimal delivery route by using the NSGA non-dominated sorting genetic algorithm, it performs clustering according to the geographical locations of the pickup points and delivery points of the orders by using the DBSCAN clustering method, so that the orders in the same clustering cluster are delivered by one vehicle.

[0068] In some embodiments of the present invention, the optimal delivery route optimization unit further includes:

[0069] A loading engine module, which is used to simulate the three-dimensional loading of cargo, record the available space and the storage coordinates and order of the cargo, and the loading process satisfies various constraint conditions in the model;

[0070] A feasibility inspection module, which is used to check the order of the pickup points and delivery points of the route, and check whether the order is completed in combination with the loading engine, and regards the routes that all meet as feasible routes.

[0071] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the multi-stop path planning method and system based on the urban distribution scenario proposed by the present invention, the traditional path planning model and the three-dimensional loading model are associated and modeled and integrated to establish a multi-stop path planning and 3D loading integration model with the total cost, time penalty function, and total vehicle full load rate as multiple objectives. Path constraints such as road restrictions, store opening hours, and driver fatigue driving are added, as well as loading constraints such as large items not pressing on small items, heavy items not pressing on light items, limited load volume, and first-in-last-out, to obtain an integrated optimization result of VRP and 3D-CLP, realizing an equilibrium and optimal decision on vehicle paths and cargo loading methods, improving the safety and loading rate of the loading and transportation system, and significantly reducing the cargo damage rate.

[0072] After reading the detailed description of the embodiments of the present invention in conjunction with the accompanying drawings, other features and advantages of the present invention will become clearer. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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 be obtained based on these drawings.

[0074] Figure 1 Schematic diagram of the steps of the multi-stop path planning method based on the urban distribution scenario proposed by the present invention;

[0075] Figure 2 Schematic diagram of the establishment of the three-dimensional loading coordinate system in the present invention;

[0076] Figure 3 Schematic diagram of the steps of the embodiment of the multi-stop path planning method based on the urban distribution scenario proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0077] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0078] The purpose of this invention is to construct a multi-point path planning and assembly integration model (referred to as the loading integration model) designed specifically for urban distribution scenarios, aiming to simultaneously consider the impact of cargo loading sequence and vehicle pickup path on key operational factors such as vehicle loading rate and transportation cost, so as to obtain a result of VRP and 3D-CLP integrated optimization to achieve a balanced optimal decision for vehicle path and cargo loading method.

[0079] In the process of building this model, the following key issues were considered in depth:

[0080] (1) Optimal path planning.

[0081] In the specific implementation of multi-point route planning, it is necessary to take into account the entire distance from the initial starting point before the vehicle receives the order to the pickup point and then to the delivery point. It is necessary to adopt a reasonable route planning strategy to ensure that the optimal "carpooling combination" between orders is sought without violating vehicle restrictions, load constraints, etc., and through refined route planning, transportation efficiency can be improved and operating costs can be reduced.

[0082] (2) Cargo loading strategy.

[0083] Based on the analysis of the principles of cargo loading and placement, in the practical application of vehicle loading problems, the optimization problem is described as: Under certain constraints, a reasonable loading strategy is used to ensure that the cargo specified in the distribution task can be efficiently distributed to the most suitable vehicle model. This process aims to maximize the overall transportation efficiency of the vehicle while meeting the distribution needs and achieve economic optimization of vehicle model selection. Specifically, factors such as the size, weight, and fragility of the cargo, as well as the spatial layout and loading order of the vehicle compartment, need to be considered to ensure the rationality and efficiency of the loading plan.

[0084] The cargo loading and placement principles here include loading principles and placement order principles:

[0085] The loading principles include:

[0086] 1) Cargo loading sequence: For the problem of non-full load multi-point distribution, in fact, in order to reduce the possible reloading process of cargo at the intermediate unloading point, the cargo loading sequence should be closely related to the entire distribution plan. The basic principles are "unload first and load later", "heavy goods do not press light goods", and "large goods do not press small goods".

[0087] 2) Loading the same batch of goods on the same vehicle: For the same batch of goods with the same destination, they should be loaded on the same vehicle as much as possible and placed as centrally as possible to facilitate loading, unloading and management.

[0088] 3) Mixed loading constraint problem: That is, there may also be "compatibility" between goods in actual applications. And it cannot be simply considered as "can they be loaded on the same vehicle", but should be considered as whether they can be placed adjacent to each other or even overlapped, even if they are loaded on the same vehicle.

[0089] 4) General "compatibility" constraint: For goods with "compatibility" constraint requirements, if they are allowed to be transported on the same vehicle, there should be at least a certain space interval between them.

[0090] 5) Center of gravity balance problem: Unreasonable distribution of goods will cause the center of gravity of the carriage to shift. Therefore, the placement position of goods in the carriage will affect the safety performance of vehicle driving. The self-weight of the goods and the straight-line distance to the center of the vehicle will cause different torques for each good. Unbalanced forces will generate torque differences in the horizontal or vertical directions, thus causing the shift of the center of gravity. Further consequences are: affected by gravity, the two sides of the carriage produce inclination angles, resulting in safety hazards such as inconsistent tire wear and route deviation during vehicle driving. Ensure that the goods are evenly and stably distributed on the vehicle, avoid overloading, partial loading, and concentrated weight phenomena, and keep the total center of gravity of the goods and the vehicle within a reasonable range to ensure safety and efficiency during transportation. When the goods are loaded, the offset between the integrated center of gravity of the goods and the carriage and the physical center of gravity of the carriage shall not exceed the maximum value allowed for the offset of the vehicle's center of gravity.

[0091] 6) Goods overlapping problem: Only stacking of the same category, arbitrary stacking of any category, and non-allowance of stacking. When goods are placed overlapping, they cannot exceed the limit of the maximum allowable overlapping layers (when different goods are overlapped, it is limited by the minimum allowable overlapping layers).

[0092] 7) Restrictions on special goods: For example, food cannot be mixed with toxic goods and chemical products. There are objective requirements for the loading methods of fragile goods and dangerous goods. In addition, some goods may also have certain restrictions on the placement method during loading. For example, goods such as refrigerators and air conditioners are not allowed to be inverted, and televisions, computer monitors, etc. cannot only be inverted, but even have certain orientation requirements for a specific surface in the placement method. In addition, due to special reasons such as the "incompatibility" between goods, such as food and washing products cannot be placed adjacent to each other, or heavier goods cannot be stacked on top of fragile goods, etc., also need to be considered.

[0093] 8) Selection of vehicle type. During the entire process of loading goods, on the basis of meeting the goods placement conditions, after finding a more suitable placement method and loading sequence, more goods can be loaded into the cargo compartment of a more economical means of transportation (such as a smaller vehicle type), taking into account the vehicle passing conditions and traffic control of the vehicle type, so as to improve the efficiency of the logistics distribution system.

[0094] The placement sequence principles include:

[0095] 1) Try to place the goods in sequence from left to right along the left side to the right side of the carriage (the positive y-axis direction).

[0096] 2) Place the goods layer by layer. For the same batch of goods (especially those of the same variety and specification), try to use the stacking method of overlapping placement as much as possible. In other cases, also place them along the direction from the lower bottom surface to the upper bottom surface (the positive z-axis direction).

[0097] 3) Try to place the goods in sequence from the front to the back, from the front end (near the cab) to the back end (the rear of the vehicle) of the vehicle body (the x-axis direction).

[0098] (3) Vehicle operation safety.

[0099] To ensure that the model can be applied to solve real problems, it is necessary to consider the urban distribution and transportation scenarios realistically. For example, during the vehicle operation process, the loading method of the goods must ensure that the center of gravity of the carriage meets the requirements of safe driving, and at the same time follow the basic safety principles such as large goods not pressing on small ones and heavy goods not pressing on light ones. In addition, due to the complexity of the urban road conditions, in order to reduce the accident risk, when dispatching orders to vehicles, it is also necessary to fully consider the danger of driver fatigue driving to ensure the safety and reliability of the transportation process.

[0100] (4) Truck scheduling.

[0101] Considering the uniqueness of transportation in the urban distribution 2B scenario, more detailed planning should be carried out for the operation time of the fleet. This is mainly aimed at various traffic restriction rules for trucks of different specifications in the city, as well as the personalized requirements of B-end customers for the receiving time of goods in the urban distribution scenario. Through reasonable scheduling, ensure that the trucks can arrive at the designated pick-up points and delivery points on time on the premise of meeting the traffic restriction rules, and meet the distribution needs of customers.

[0102] (5) Calling of non-standard vehicle models.

[0103] Considering that it is difficult to arrange the loading of non-standard vehicle models according to the regulations of conventional vehicle models when participating in urban distribution transportation tasks, in actual applications, the intuitive judgment of the available capacity of the vehicle by drivers and loading staff is often relied on, and there are often violations of not considering the vehicle weight limit in order to "fill the vehicle". The consequence is that it may lead to overloading, over-limit, rear rollover, side rollover or even road collapse during the vehicle driving process.

[0104] Based on the above analysis, the present invention abandons the traditional path model that plans the specific transport capacity configuration according to the general models and specification standards of freight trucks, and innovatively adopts a more refined parameter setting method, that is, taking the three dimensions of the length, width, and height of the carriage, as well as the corresponding loadable volume and limited load weight inside the carriage, as the core parameters for defining the vehicle capacity. Specifically, when calculating the transport capacity of the vehicle, the model constructed by the present invention fully takes into account the two key factors of loading efficiency and weight limit, and strives to minimize the idle space in the carriage on the premise of strictly complying with the weight limit regulations, so as to effectively improve the loading rate of the vehicle.

[0105] In the process of formulating the optimization strategy for the goods loading method, based on the special requirements of the urban distribution scenario mentioned above, the present invention widely incorporates various constraint conditions, such as the load capacity of the vehicle, spatial layout, and characteristics of the goods, comprehensively analyzes and compares the actual loading potential of different vehicle models in combination with the intelligent algorithm model, and then accurately formulates the most suitable loading plan to ensure the safety and stability of the goods during transportation.

[0106] In addition, the present invention also focuses on the control of transportation costs, and strives to seek the most economically viable solution while ensuring the transportation quality and safety. It can not only effectively promote the rational allocation and efficient utilization of vehicle resources in the goods distribution process, but also significantly improve the overall transportation efficiency, greatly reduce the empty load rate and resource waste phenomenon. This efficient utilization mode of vehicle resources not only helps to reduce transportation costs, but also significantly improves the overall economic benefits of the enterprise, thus creating a broader development space and profit space for the enterprise.

[0107] Before constructing the integrated model of multi-string point path planning and 3D loading, the following assumptions are made for the model:

[0108] 1. Assume that the parking places of the vehicles are a finite number of known fixed sites. When the vehicle completes the day's tasks or rests, it will return to its corresponding parking place, and the coordinate positions of each fixed site are known.

[0109] 2. Assume that the transportation scenario of this model conforms to the urban distribution distance limit, that is, there is no vehicle change during the transportation process.

[0110] 3. There are multiple pick-up points and multiple delivery points, without transfer points; and the coordinate positions of each node (pick-up point and delivery point) are known.

[0111] 4. A vehicle can go to multiple pick-up points to pick up goods, that is, to achieve the less-than-truckload business, and can also go to multiple delivery points to deliver goods.

[0112] 5. As much as possible, the goods of the same customer are transported by full truckload.

[0113] 6. The starting site information corresponding to all available vehicles is known. When receiving an order, the vehicle departs from the starting point, picks up the goods at the pick-up point, and delivers them to the delivery point; and when completing one order task and continuing to receive orders, the last delivery point is used as the next starting point information.

[0114] 7. Simplify the loading and unloading time of the truck when loading at the warehouse and unloading at the B-end, that is, it is considered that the time required for each vehicle type to load and unload the same kind of goods is the same, and the truck does not need to wait for loading and unloading operations at the warehouse or the B-end.

[0115] 8. The shapes of all goods for stowage are regular.

[0116] 9. The load weight and volume of a single piece of goods shall not exceed the maximum load and volume of the vehicle.

[0117] 10. Assume that the specifications and quantities of available vehicles are known.

[0118] 11. The time limit requirements, pick-up points, delivery points, types of goods, regulations and quantity information for each order are known.

[0119] 12. The time limit requirements for orders are divided into three cases: same-day delivery, next-day delivery, and arrival within the specified time without distinction between morning and evening. The time limit requirements for all goods in each order are the same. Penalty costs are incurred for arriving after the customer's expected arrival time (soft time window), and arrival must be before the latest arrival time (hard time window).

[0120] 13. Assume that the vehicle restriction regulations for all road sections are divided into three time periods, namely morning and evening rush hours and normal time periods. Define the morning rush hour as T0 - T1 (for example, 7:00 - 9:00), the evening rush hour as T2 - T3, and the time period in between as the normal time period T1 - T2.

[0121] 14. Assume that the goods can be picked up immediately at the pick-up point after receiving the order, that is, there is no shortage of goods at the pick-up point.

[0122] 15. Assume that the vehicle less-than-truckload logistics is in a continuous state in terms of time.

[0123] 16. Define the limit of a driver's fatigue driving as continuous working for 4 hours or more. A driver whose continuous working time exceeds this limit will be stopped from receiving orders and return to the parking lot to rest.

[0124] Based on the above, the objective function constructed by the present invention includes three parts: the total cost during vehicle distribution (including vehicle fixed costs and variable costs), the time penalty function for violating the order timeliness requirements, and the total vehicle full load rate (including volume full load rate and load full load rate), which will be described one by one below

[0125] 1. Construct the total cost objective function of vehicle distribution with the goal of minimizing vehicle fixed costs and variable costs.

[0126] In this part of the total cost of vehicle distribution, the fixed vehicle costs are mainly the fixed costs for vehicle startup. If there are any, they also include a small part of vehicle wear and tear, driver salaries, etc.; the variable costs are mainly the fuel consumption generated during vehicle driving. Its functional expression is:

[0127]

[0128] 2. Construct a time penalty function with the goal of minimizing the cost of violating the order timeliness.

[0129] To represent the constraint on order timeliness, the present invention sets up a penalty function for time in the model, that is, a cost needs to be paid when the goods arrival time exceeds the customer's expected arrival time, and the goods arrival time is not allowed to be later than the latest arrival time. The penalty function and the objective function are respectively expressed as:

[0130]

[0131]

[0132] 3. Construct a total vehicle full-load rate function with the goal of maximizing the volume full-load rate and the load full-load rate.

[0133] Since this model considers the problem of pick-up and delivery at multiple stops, that is, there is a situation where when a vehicle departs from a certain node, the carriage is loaded with goods from different pick-up points. Therefore, in order to conveniently represent the specific loading situation of the vehicle when driving between any two nodes, a binary variable A ink is introduced, which is used to represent whether the goods taken out from node i are loaded in the carriage when vehicle k leaves node n. If so, it is 1; otherwise, it is 0. It is expressed as:

[0134] A ink = sgn[(a nk - a ik )(a (N+i)k - a nk )];

[0135] Then, the objective function related to the vehicle full-load rate is defined from two aspects: the volume full-load rate and the load full-load rate. The functional expression is as follows:

[0136]

[0137] The constraint conditions include:

[0138] 1. Vehicle routing problem constraints:

[0139] P ∩ D = 0; (1)

[0140]

[0141] Among them, constraint (1) means there is no transfer point during transportation and loading; constraint (2) means that all goods starting from a pickup point are delivered by one vehicle; constraint (3) means that the vehicle will not return from the delivery point to the pickup point; constraints (4-5) mean that any vehicle will start from a certain point and then drive to the next point after arriving at a certain point; constraint (6) means that an order can only be assigned to a vehicle when the route of the vehicle for the order is not restricted; constraint (7) means that any vehicle must visit the pickup point of the order before visiting the corresponding delivery point; constraint (8) means the arrival time of the vehicle at the node; constraint (9) means the waiting time for loading and unloading of the vehicle at the node; constraint (10) means that the departure time of the vehicle from the starting point must be later than the order placement time of the order; constraint (11) means the departure time of the vehicle from the node; constraint (12) means the traffic restriction constraint that vehicle k should comply with from node i to node j; constraint (13) means that the continuous working time of the driver shall not exceed the fatigue driving limit.

[0142] 2. Constraints of the three-dimensional loading problem:

[0143] Taking the lower left corner of the carriage as the origin, establish a three-dimensional coordinate system as shown in Figure 2 . Let m (the m-th piece of goods loaded from node i) be the coordinates of the lower left rear vertex of goods i on vehicle k, Let m be the coordinates of the upper right front vertex of goods i on vehicle k, Let m be the length, width and height of goods i.

[0144] Then when vehicle k finishes loading at node n, the integrated center of gravity coordinates U(Φ n,k,x , Φ n,k,y , Φ n,k,z ) can be expressed as:

[0145]

[0146] And for the convenience of expressing the state of goods i p and goods j q in the carriage of vehicle k at the same time, introduce a binary variable If goods i p and goods j q are in the carriage of vehicle k at the same time, it is 1, otherwise it is 0. It is expressed as:

[0147]

[0148] Constraints:

[0149]

[0150]

[0151] For all i p ≠ j q , one of each of the following three groups of expressions must be satisfied:

[0152]

[0153] For all the following two expressions must be satisfied:

[0154]

[0155] If cargo i p is unloaded before cargo j q , the following group of expressions must be satisfied:

[0156]

[0157] Among them, constraints (14 - 16) indicate that all cargos should be placed in parallel in the carriage; constraints (17 - 19) indicate that after the cargos are loaded, the center of gravity in the carriage should be within the range required for safe driving; constraint (20) indicates that the sum of the weights of the cargos loaded by the vehicle on any section of the road should be less than the rated load of the vehicle; constraint (21) indicates that the sum of the volumes of the cargos loaded by the vehicle on any section of the road should be less than the rated volume of the vehicle; constraints (22 - 24) indicate the cargo placement range constraints, that is, the cargo coordinates cannot exceed the inner edge coordinates of the carriage; constraints (25 - 27) indicate that the spatial positions of any two cargos placed in the same carriage cannot overlap; constraint (28) indicates that the heavier cargo cannot be placed on top of the lighter cargo; constraint (29) indicates that the cargo with a larger bottom area cannot be placed on top of the cargo with a larger bottom area; constraint (30) indicates the first-in, last-out constraint for cargos.

[0158] The above parameters are explained as shown in the following table:

[0159] Table 1 Parameter Explanation

[0160]

[0161] Due to the particularity of urban distribution orders, with a large number of orders, scattered locations, and strong timeliness, and the model is proposed with the total driving cost, time penalty, and vehicle loading rate as the objectives, combining path and 3D loading integration for optimization. In order to be able to complete the order processing and vehicle dispatching smoothly and efficiently, the present invention uses a two-stage method for multi-objective optimization implementation. As Figure 3As shown in the figure, the present invention first clusters orders using DBSCAN (Density-Based Spatial Clustering of Applications with Noise) according to the geographical locations of the pick-up points and delivery points of the orders. Orders within the same cluster are delivered by the same vehicle. Then, the NSGA (Non-dominated Sorting Genetic Algorithm) non-dominated sorting genetic algorithm is used to optimize the routes for each vehicle. NSGA is a multi-objective genetic algorithm mainly used to solve multi-objective optimization problems. Compared with the traditional GA genetic algorithm, by introducing strategies such as non-dominated sorting, crowding degree calculation, and tournament selection, it can quickly make a choice for multi-objective optimization problems, find the Pareto front solutions of multi-objective optimization problems, improve the efficiency of the algorithm and the quality of the results, and can well maintain the diversity of solutions. Finally, by solving the optimal routes of the vehicles for each cluster of orders, the optimal results of the final cargo loading and urban distribution integration are synthesized.

[0162] Specifically, in the data layer stage, the present invention first defines entity classes for vehicles, cargo, and orders, laying a data foundation for the subsequent algorithms. Then, it designs a simulated data generation function to generate corresponding vehicle, cargo, and order data for use in subsequent algorithms. In the preprocessing stage, it first performs DBSCAN clustering on the orders according to geographical locations, clusters the orders, and arranges vehicles to complete the delivery of each cluster of orders. In the core algorithm stage, it designs a loading engine module to simulate the three-dimensional packing of cargo, record the available space, storage coordinates, and order of the cargo. The loading process needs to meet various constraints in the model. Then, it designs a feasibility inspection module to check the pick-up and delivery point order of the route and, in combination with the loading engine, check whether the order is completed. Only the routes that meet all the requirements can be regarded as feasible routes. Then, it designs the NSGA non-dominated sorting genetic algorithm to generate an effective population through operations such as initial population generation, crossover, and mutation. Then, according to the model, three objective functions are designed, namely the transportation cost, time window penalty, and dynamic loading rate function. Through the elitist strategy: judging the dominance relationship, tournament selection, non-dominated sorting, and crowding degree calculation, the Pareto front solutions are quickly calculated. Then, it designs the main program function to call the algorithm for solution, find the optimal routes for each cluster of orders, integrate them into the final optimal result, and visualize it. The following uses a specific embodiment to elaborate in detail on the implementation and effects of the multi-string point route planning method proposed by the present invention.

[0163] The operation background content includes:

[0164] 1. The common truck regulations are shown in Table 2 below:

[0165] Table 2 Common Truck Regulations Table

[0166]

[0167] 2. The regulations for the soft and hard time windows of common urban distribution orders are shown in Table 3 as follows: Table 3 Regulations for the soft and hard time windows of common urban distribution orders

[0168]

[0169] 3. The node coordinate (example) information is shown in Table 4 as follows: Table 4 Node coordinate (example) information

[0170]

[0171] 4. The selected values of the known parameters are shown in Table 5 as follows:

[0172] Table 5 Selected values of the known parameters

[0173]

[0174] 5. The classification information of some common goods is shown in Table 6 as follows: Table 6 Information on common types of goods

[0175]

[0176] In this implementation, 6 pick-up points, 10 corresponding delivery points, and the corresponding basic information such as longitude and latitude positions, vehicle information, and cargo sizes in the distribution data table are selected as input data.

[0177] The main parameter settings of the large neighborhood algorithm based on the nearest neighbor are as follows: the number of iterations is 200 times, and the convergence graph of the algorithm is shown in the figure.

[0178] Finally, three vehicles are selected for distribution according to the distribution information of 10 customers, and three corresponding distribution routes for multi-point pick-up and multi-point delivery are formed as shown in Table 7. The cargo loading conditions of the corresponding vehicles and the comparative analysis of the transportation costs before and after optimization are shown in Tables 8 and 9 respectively.

[0179] Table 8 Vehicle distribution route table

[0180]

[0181] Table 9 Vehicle cargo loading conditions

[0182]

[0183] Table 9 Analysis of vehicle transportation costs before and after optimization

[0184]

[0185] After considering more actual packing constraints, such as vehicle center of gravity in the three-dimensional packing model and various loading and unloading principles (such as the constraint that large items do not press on small ones and heavy items do not press on light ones), and solving the two problems of vehicle routing and vehicle loading simultaneously to handle the integrated loading and distribution problem. Finally, the total distance of the vehicle transportation form before and after optimization is optimized by 9%, the average maximum loading rate is increased by 18.4%, and the total cost is correspondingly reduced by 10.5%. Comprehensive optimization in multiple aspects and multiple indicators is achieved.

[0186] Based on the above, the present invention proposes a multi-drop point path planning system based on the urban distribution scenario to implement the above-mentioned multi-drop point path planning method. The system includes:

[0187] The total cost construction unit is used to construct the total cost objective function of the vehicle distribution process with the goal of minimizing the fixed cost and variable cost of the vehicle.

[0188] The time penalty function construction unit is used to construct the time penalty function with the goal of minimizing the cost of violating the order timeliness.

[0189] The vehicle total full load rate function construction unit is used to construct the vehicle total full load rate function with the goal of maximizing the volume full load rate and the load full load rate.

[0190] The optimal distribution path solving unit is used to solve the total cost objective function, the time penalty function and the vehicle total full load rate function based on the NSGA non-dominated sorting genetic algorithm to obtain the optimal distribution path for multi-point pick-up and multi-point delivery.

[0191] The path planning method of the specific system has been described in detail above and will not be elaborated here.

[0192] It should be noted that in the specific implementation process, the above control part can be realized by a processor in the form of hardware executing computer execution instructions in software form stored in the memory, which will not be elaborated here, and the programs corresponding to the actions executed by the above control circuit can all be stored in the computer-readable storage medium of the system in software form, so as to facilitate the processor to call and execute the operations corresponding to each module above.

[0193] The computer-readable storage medium in the above text may include volatile memory, such as random access memory; it may also include non-volatile memory, such as read-only memory, flash memory, hard disk or solid-state drive; it may also include a combination of the above types of memory.

[0194] The processor mentioned above may also be a general term for multiple processing elements. For example, the processor may be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, field programmable gate arrays, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc., and may also be a dedicated processor.

[0195] It should be noted that the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions, or substitutions made by those of ordinary skill in the art within the scope of the essence of the present invention should also fall within the protection scope of the present invention.

Claims

1. A multi-string point path planning method based on the urban distribution scenario, characterized in that Including: Construct a total cost objective function for the vehicle distribution process with the goal of minimizing the fixed and variable costs of the vehicle: Among them, D k is the fixed cost of calling vehicle k, and x ijk indicates whether vehicle k travels from node i to node j. If so is 1, otherwise 0; C k is the fuel cost consumed per kilometer by vehicle k; d ij represents the driving distance from node i to node j; Construct a time penalty function with the goal of minimizing the cost of violating the order timeliness: Among them, Among them, T i 1 represents the customer's expected arrival time of the order picking up goods from node i; T i 2 represents the latest arrival time of the order picking up goods from node i; p represents the penalty cost coefficient for the goods arriving later than its soft time window; Construct a total vehicle full load rate function with the goal of maximizing the volume full load rate and the load full load rate: Among them, A ink is used to indicate whether the vehicle k is loaded with the goods taken out from node i when leaving node n. If so, it is 1; otherwise, it is 0; w im represents the weight of the m-th piece of goods taken from node i; x njk indicates whether vehicle k drives from node n to node j; W k represents the rated load capacity of vehicle K; V k represents the rated volume of vehicle K; Based on the NSGA non-dominated genetic algorithm, solve the total cost objective function, the time penalty function, and the total vehicle full load rate function to obtain the optimal distribution path for multi-point pick-up and multi-point delivery.

2. The multi-point path planning method based on the urban distribution scenario according to claim 1, wherein The constraint conditions of the vehicle routing problem include: P ∩ D = 0; 0 ≤ x ijk ≤ ME ijk , i ∈ I ∪ P ∪ D, j ∈ P ∪ D, k ∈ K; a ik <a (N+i)k , i ∈ P; Among them, P = {1, 2,... N} is the set of the locations of all pick-up points in the distribution network; D = {N + 1, N + 2,... 2N} is the set of the locations of all delivery points in the distribution network; y ik indicates whether the goods at the i-th pick-up point are delivered by vehicle k. If so, it is 1; otherwise, it is 0; E ijk indicates whether vehicle k can pass through the section from node i to node j. If it can, it is 1; otherwise, it is 0; a ik is the time when vehicle k arrives at node i; r ik indicates the time when vehicle k leaves node i; y ik indicates whether the goods at the i-th pick-up point are delivered by vehicle k. If so, it is 1; otherwise, it is 0.

3. The multi-point path planning method based on the urban distribution scenario according to claim 1, wherein The three-dimensional loading constraint conditions include: For all i p ≠ j q , one of each of the following three sets of expressions must necessarily be satisfied: For all the following two equations must be satisfied: If goods i p is unloaded before goods j q is unloaded, then the following set of equations must be satisfied: Among them, represents the left rear lower vertex coordinates of the m-th piece of cargo taken by vehicle k from node i; is the cargo i m at the upper right front vertex coordinates on vehicle k; is the cargo i m length, width, and height; expresses the state of cargo i p and cargo j q in the carriage of vehicle k at the same time. If cargo i p and cargo j q are in the carriage of vehicle k at the same time, it is 1, otherwise it is 0.

4. The multi-string point path planning method based on the urban distribution scenario according to claim 1, characterized in that, Before using the NSGA non-dominated sorting genetic algorithm to solve the optimal distribution path, the method further includes: Cluster according to the geographical locations of the pick-up points and delivery points of the orders using the DBSCAN clustering method, so that the orders in the same cluster are delivered by one vehicle.

5. The multi-string point path planning method based on the urban distribution scenario according to claim 4, wherein The method further includes: Design a loading engine module to simulate the three-dimensional packing of goods, record the available space, the storage coordinates and order of the goods, and the loading process satisfies various constraint conditions in the model; Design a feasibility inspection module to check the order of the pick-up points and delivery points of the path, and combine with the loading engine to check whether the order is completed, and regard the path that satisfies both as a feasible path.

6. A multi-string point path planning system based on the urban distribution scenario, characterized in that, Including: A total cost construction unit for constructing a total cost objective function for the vehicle distribution process with the goal of minimizing the fixed and variable costs of the vehicle: Among them, D k is the fixed cost of calling vehicle k, and x ijk indicates whether vehicle k travels from node i to node j; if so is 1, otherwise 0; C k is the fuel cost consumed by vehicle k per kilometer; d ij represents the driving distance from node i to node j; A time penalty function construction unit for constructing a time penalty function with the goal of minimizing the cost of violating the order timeliness: Among them, Among them, T i 1 represents the expected arrival time of the customer of the order picking up goods from node i; T i 2 represents the latest arrival time of the order picking up goods from node i; p represents the penalty cost coefficient for the goods arriving later than its soft time window; A total vehicle full load rate function construction unit for constructing a total vehicle full load rate function with the goal of maximizing the volume full load rate and the load full load rate: Among them, A ink is used to indicate whether the vehicle k is loaded with the goods taken from node i when leaving node n. If so, it is 1; otherwise, it is 0; w im represents the weight of the m-th piece of goods taken from node i; x njk indicates whether the vehicle k drives from node n to node j; W k represents the rated carrying capacity of vehicle K; V k represents the rated volume of vehicle K; An optimal distribution path solving unit for solving the total cost objective function, the time penalty function, and the total vehicle full load rate function based on the NSGA non-dominated genetic algorithm to obtain the optimal distribution path for multi-point pick-up and multi-point delivery.

7. The multi-string point path planning system based on the urban distribution scenario according to claim 6, wherein The constraint conditions of the vehicle routing problem include: P ∩ D = 0; 0 ≤ x ijk ≤ ME ijk , i ∈ I ∪ P ∪ D, j ∈ P ∪ D, k ∈ K; a ik <a (N+i)k , i ∈ P; Among them, P = {1, 2,... N} is the set of the locations of all pickup points in the distribution network; D = {N + 1, N + 2,... 2N} is the set of the locations of all delivery points in the distribution network; y ik indicates whether the goods at the i-th pickup point are delivered by vehicle k. If so, it is 1; otherwise, it is 0; E ijk indicates whether vehicle k can pass through the section from node i to node j. If it can, it is 1; otherwise, it is 0; a ik is the time when vehicle k arrives at node i; r ik indicates the time when vehicle k leaves node i; y ik indicates whether the goods at the i-th pickup point are delivered by vehicle k. If so, it is 1; otherwise, it is 0.

8. The multi-string point path planning system based on the urban distribution scenario according to claim 6, wherein The three-dimensional loading constraint conditions include: For all i p ≠ j q , one of each of the following three sets of expressions must necessarily be satisfied: For all the following two equations must be satisfied: If cargo i p is unloaded before cargo j q is unloaded, then the following set of equations must be satisfied: Among them, represents the left rear lower vertex coordinates of the m-th piece of cargo taken by vehicle k from node i; is the cargo i m at the upper right front vertex coordinates on vehicle k; is the cargo i m length, width, and height; expresses the state of cargo i p and cargo j q simultaneously in the carriage of vehicle k. If cargo i p and cargo j q are simultaneously in the carriage of vehicle k, it is 1, otherwise it is 0.

9. The multi-string point path planning system based on the urban distribution scenario according to claim 6, characterized in that, Before using the NSGA non-dominated sorting genetic algorithm to solve the optimal distribution path, the optimal distribution path solving unit clusters according to the geographical locations of the pick-up points and delivery points of the orders using the DBSCAN clustering method, so that the orders in the same cluster are delivered by one vehicle.

10. The multi-string point path planning system based on the urban distribution scenario according to claim 9, characterized in that, The optimal distribution path optimization unit further includes: A loading engine module for simulating the three-dimensional packing of goods, recording the available space, the storage coordinates and order of the goods, and the loading process satisfies various constraint conditions in the model; A feasibility inspection module for checking the order of the pick-up points and delivery points of the path, and combining with the loading engine to check whether the order is completed, and regarding the path that satisfies both as a feasible path.

Images