Pickup optimization method and system for intelligent logistics

Abstract

The invention relates to the field of logistics, in particular to a pick-up optimization method and system for intelligent logistics, and the method comprises the steps: obtaining a to-be-processed pick-up task set, courier information and real-time road network data; in path construction of each iteration of the ant colony system, according to path pheromone concentration, comprehensive heuristic information, courier area proficiency and task urgency, calculating the transfer probability of ants from a current node # imgabs0 # to a next non-accessed node # imgabs1 #; after each iteration is completed, calculating a space-time volatilization factor according to the historical traffic efficiency and the real-time congestion index, and updating global pheromones by using the space-time volatilization factor; performing pheromone enhancement on the optimal path based on the path quality of the current iteration optimal solution; the path quality is a scalar value of the total duration of the comprehensive evaluation path, the time window satisfaction degree and the comprehensive path cost; and repeating the path construction and pheromone updating steps until a preset termination condition is met, and outputting an optimal pickup path sequence.

Description

Technical Field

[0001] The present invention relates to the field of logistics, and specifically to a pickup optimization method and system for smart logistics. Background Art

[0002] The efficiency of last-mile delivery is a core indicator for measuring the quality of logistics services and controlling operating costs. Traditional route planning methods are mostly static, relying primarily on fixed map data to find the shortest distance or fastest route solution, making it difficult to cope with the complex dynamic changes in the real world. For example, urban traffic conditions are affected by multiple factors such as time, weather, and emergencies. Real-time traffic congestion and parking difficulties in specific areas pose a huge challenge to the actual execution efficiency of routes. In addition, traditional methods often treat all couriers and tasks as homogeneous processing units, ignoring the differences in individual couriers' proficiency in different areas, as well as the customer time windows and urgency requirements of different pickup tasks. This makes the planned route solutions often deviate from actual operations and lack flexibility and practical guidance.

[0003] Ant Colony Optimization (ACO), with its positive feedback, distributed computing, and strong robustness, has shown great potential for solving complex combinatorial optimization problems and is an effective tool for dynamic route planning. However, when applied to complex logistics scenarios, the basic ACO still suffers from slow convergence and a tendency to become trapped in local optima. In particular, traditional models fail to effectively integrate spatiotemporal traffic data in their pheromone update mechanisms, and they also fail to fully consider multi-dimensional dynamic factors such as task attributes and operator experience in the design of heuristic information for route selection. Therefore, improving the ACO so that it can comprehensively process multi-dimensional information such as real-time road conditions, task priorities, and operator proficiency to generate more efficient, accurate, and reliable pickup routes is a technical challenge that urgently needs to be overcome in the field of smart logistics technology. Summary of the Invention

[0004] To address the issue of wasted time in dispatching and picking up parcels in logistics, in a first aspect of the present invention, a method for optimizing parcel pickup for smart logistics is provided, comprising the following steps: S1: Obtain a collection of pending pickup tasks, courier information, and real-time road network data. The pickup task collection includes the location, customer time window, and task urgency of each task. The courier information includes proficiency in each area. The real-time road network data includes the road network topology, predicted travel time between nodes, real-time congestion index, historical traffic efficiency, regional parking difficulty, and estimated waiting time at the target pickup point. S2, in each iteration of the path construction of the ant colony system, the courier’s path from the current node is calculated based on the path pheromone concentration, comprehensive heuristic information, courier’s regional proficiency and task urgency. Go to the next unvisited node The transition probability is , where the comprehensive heuristic information is the inverse of the comprehensive path cost, which is a scalar value that comprehensively evaluates the predicted travel time between nodes, the difficulty of parking in the area, and the estimated waiting time at the target pickup point. S3, after each iteration, the spatiotemporal volatility factor is calculated based on the historical traffic efficiency and real-time congestion index, and the global pheromone is updated with the spatiotemporal volatility factor. Based on the path quality of the optimal solution of the current iteration, the optimal path is enhanced with pheromones. The path quality is a scalar value that comprehensively evaluates the total path duration, time window satisfaction, and comprehensive path cost. S4, repeat the path construction and pheromone update steps until the preset termination condition is met, and output the optimal pickup path sequence.

[0005] During the path selection phase, the present invention constructs a more realistic comprehensive path cost by comprehensively evaluating the predicted travel time between nodes, the difficulty of parking in the area, and the estimated waiting time at the target pickup point, making the basis for path decision-making more comprehensive and accurate. At the same time, the courier's proficiency in a specific area and the urgency of the task itself are introduced into the probability model of path construction, so that the generated path plan takes into account the individual capabilities of the executor and the specific requirements of the task, and realizes refined scheduling. The volatility of pheromones is set by using historical traffic efficiency and real-time congestion index, and the influence of high-quality solutions is enhanced based on a comprehensive evaluation of the total path duration, time window satisfaction, and comprehensive cost, thereby improving the global optimization ability and convergence effect of path planning, and being able to output a pickup path sequence with shorter total time, better service quality, and higher reliability in actual execution.

[0006] Optionally, calculate the courier from the current node Go to the next unvisited node The transition probability , calculated using the following formula: ; in, for Time Path The pheromone concentration on To provide comprehensive inspiration information, For couriers For Node Proficiency in the area For nodes The urgency of the task; 、 、 、 is the preset non-negative weight parameter; the sum of the denominator part is for the courier All accessible next nodes Perform calculations.

[0007] Optionally, the calculation process of the comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The quantified value of parking difficulty in the area and the estimated waiting time, where the quantified value of parking difficulty is proportional to the parking difficulty; The quantified value of parking difficulty is multiplied by the estimated waiting time, and then summed with the predicted travel time to obtain the comprehensive path cost; The inverse of the comprehensive path cost is used as the comprehensive heuristic information.

[0008] It comprehensively considers the difficulty of driving and parking and the waiting time, and is more in line with actual scenarios.

[0009] Optionally, the size of the spatiotemporal volatility factor is positively correlated with the real-time congestion index and negatively correlated with the historical traffic efficiency.

[0010] Combine real-time congestion with historical traffic efficiency to dynamically reflect changes in road conditions.

[0011] Optionally, pheromone enhancement is performed on the optimal path, comprising the following steps: Calculate the path quality of the current iterative optimal solution. The path quality is a function value that is proportional to the time window satisfaction and inversely proportional to the total path duration and the comprehensive path cost. Apply an additional pheromone increment proportional to the path quality to the edges on the optimal path.

[0012] Path quality is related to time window satisfaction, total path duration, and comprehensive path cost, which not only emphasizes the time window compliance but also takes into account efficiency and cost.

[0013] Optionally, the urgency of the task is inversely proportional to the time limit required by the customer to complete the task; the courier's proficiency in each area is directly proportional to the courier's familiarity with the area where the target node is located.

[0014] The combination of the urgency of the task and the proficiency of the courier is more in line with the actual pickup scenario.

[0015] In a second aspect of the present invention, a pickup optimization system for smart logistics is provided, comprising the following modules: The data collection module is used to obtain a set of pending pickup tasks, courier information, and real-time road network data. The pickup task set includes the location, customer time window, and task urgency of each task. The courier information includes proficiency in each area. The real-time road network data includes the road network topology, predicted travel time between nodes, real-time congestion index, historical traffic efficiency, regional parking difficulty, and estimated waiting time at the target pickup point. The transfer probability calculation module is used to calculate the courier's probability of moving from the current node to the next node in the path construction of each iteration of the ant colony system based on the path pheromone concentration, comprehensive heuristic information, courier's regional proficiency and task urgency. Go to the next unvisited node The transition probability is , where the comprehensive heuristic information is the inverse of the comprehensive path cost, which is a scalar value that comprehensively evaluates the predicted travel time between nodes, the difficulty of regional parking, and the estimated waiting time of the target pickup point. The update module is used to calculate the spatiotemporal volatility factor based on the historical traffic efficiency and real-time congestion index after each iteration, and use the spatiotemporal volatility factor to update the global pheromone. Based on the path quality of the current iterative optimal solution, the optimal path is enhanced with pheromones. The path quality is a scalar value that comprehensively evaluates the total path duration, time window satisfaction, and comprehensive path cost. The path output module is used to repeat the path construction and pheromone update steps until the preset termination conditions are met and output the optimal pickup path sequence.

[0016] Optionally, calculate the courier from the current node Go to the next unvisited node The transition probability , calculated using the following formula: ; in, for Time Path The pheromone concentration on To provide comprehensive inspiration information, For couriers For Node Proficiency in the area For nodes The urgency of the task; 、 、 、 is the preset non-negative weight parameter; the sum of the denominator part is for the courier All accessible next nodes Perform calculations.

[0017] Optionally, the calculation process of the comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The quantified value of parking difficulty in the area and the estimated waiting time, where the quantified value of parking difficulty is proportional to the parking difficulty; The quantified value of parking difficulty is multiplied by the estimated waiting time, and then summed with the predicted travel time to obtain the comprehensive path cost; The inverse of the comprehensive path cost is used as the comprehensive heuristic information.

[0018] Optionally, the size of the spatiotemporal volatility factor is positively correlated with the real-time congestion index and negatively correlated with the historical traffic efficiency.

[0019] Optionally, pheromone enhancement is performed on the optimal path, comprising the following steps: Calculate the path quality of the current iterative optimal solution. The path quality is a function value that is proportional to the time window satisfaction and inversely proportional to the total path duration and the comprehensive path cost. Apply an additional pheromone increment proportional to the path quality to the edges on the optimal path.

[0020] Optionally, the urgency of the task is inversely proportional to the time limit required by the customer to complete the task; the courier's proficiency in each area is directly proportional to the courier's familiarity with the area where the target node is located.

[0021] The beneficial effect is that during the path selection stage, the present invention constructs a more realistic comprehensive path cost by comprehensively evaluating the predicted travel time between nodes, the difficulty of parking in the area, and the estimated waiting time at the target pickup point, making the basis for path decision-making more comprehensive and accurate. At the same time, the courier's proficiency in a specific area and the urgency of the task itself are introduced into the probability model of path construction, so that the generated path plan takes into account the individual capabilities of the executor and the specific requirements of the task, and realizes refined scheduling. The volatility of pheromones is set by using historical traffic efficiency and real-time congestion index, and the influence of high-quality solutions is enhanced based on a comprehensive evaluation of the total path duration, time window satisfaction, and comprehensive cost, thereby improving the global optimization ability and convergence effect of path planning, and being able to output a pickup path sequence with shorter total time, better service quality, and higher reliability in actual execution. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A flowchart of a pickup optimization method for smart logistics; Figure 2 For the current node Probability graph of transition to the next node. DETAILED DESCRIPTION

[0023] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Those skilled in the art should know that the embodiments described below are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0024] Embodiments of the method for optimizing pickup for smart logistics provided by the present invention: like Figure 1 and Figure 2 As shown in FIG, the pickup optimization method for smart logistics includes the following steps: S1, obtain the set of pending pickup tasks, courier information and real-time road network data; the pickup task set includes the location of each task, customer time window and task urgency; the courier information includes proficiency in each area; the real-time road network data includes road network topology, predicted travel time between nodes, real-time congestion index, historical traffic efficiency, regional parking difficulty and estimated waiting time at the target pickup point.

[0025] Specifically, the required information is obtained by connecting data with the logistics order management system OMS, the API interface of the electronic map service provider (such as Amap or Baidu Map), and the courier performance management system.

[0026] Pickup tasks are stored in the form of a data structure list. Each task contains the customer ID, longitude and latitude coordinates, the earliest and latest allowed pickup time, and a task urgency from 1 to 5.

[0027] The urgency of a task is inversely proportional to the client's deadline for completion. This embodiment converts the abstract concept of urgency into a concrete numerical value that can be used in mathematical operations, enabling quantitative perception and response to the priorities of different tasks. This inverse relationship ensures that tasks closer to their deadlines receive higher priority. The urgency of a task directly influences the probability calculation for path selection, ensuring that high-priority tasks receive priority in planning.

[0028] For example, a customer may request an expedited delivery within an hour with 60 minutes remaining. This might be assigned an urgency of 5.0. Another example is a standard delivery requesting delivery before 6 p.m., with 360 minutes remaining at 12 p.m., and an urgency of only 1.2. Another example is a next-day delivery request with over 1,000 minutes remaining. This urgency might be as low as 0.5. During route selection, task nodes with higher urgency are assigned higher probability weights, guiding couriers to prioritize these expedited deliveries.

[0029] A courier's proficiency in each grid area is quantified as a value between 0 and 2, calculated based on their average historical pickup and delivery efficiency in that area. A courier's familiarity with an area directly impacts their efficiency in finding specific buildings, parking, and communicating with customers. Historical delivery data, such as average delivery time and success rate, allows us to generate a proficiency profile for each courier in each area and convert it into a numerical factor. This allows route planning to optimally match people with tasks.

[0030] For example, courier Zhang San has been delivering in Community A for a long time, taking an average of only 3 minutes per order, lower than the average of 7 minutes. Therefore, his proficiency factor in Community A is given a higher value, such as 1.8. Meanwhile, new employee Li Si is visiting the community for the first time and has no historical data, so he is given a baseline or lower value, such as 0.9. When planning routes for Zhang San and Li Si, even if all other conditions are the same, a route that includes a node in Community A will be much more attractive to Zhang San than to Li Si, thus achieving the most efficient dispatch.

[0031] Real-time road network data is retrieved through API calls, forming a graph structure where nodes represent mission points or intersections, and edge weights represent predicted travel times. Regional parking difficulty and estimated wait times are obtained by querying an internal historical database or by using a pre-trained gradient boosting decision tree (GBDT) model.

[0032] S2, in each iteration of the path construction of the ant colony system, the courier’s path from the current node is calculated based on the path pheromone concentration, comprehensive heuristic information, courier’s regional proficiency and task urgency. Go to the next unvisited node The transition probability of the integrated heuristic information is the inverse of the integrated path cost, and the integrated path cost is a scalar value of the predicted travel time between integrated nodes, the difficulty of regional parking, and the estimated waiting time of the target pickup point.

[0033] Specifically, Represents a courier who starts from the current node Go to the next optional unvisited node The transition probability is calculated by a state transition formula. In an optional embodiment, the numerator of the state transition formula is the product of four terms, namely, the road section Pheromone concentration on of Power, comprehensive heuristic information of Power, courier For Node Proficiency in the region of Power and Node The urgency of the task of Power. Among them, the comprehensive heuristic information is the comprehensive path cost The reciprocal of It is a comprehensive node arrive Predicted travel time and nodes The scalar value of parking difficulty and estimated waiting time. The denominator is the sum of the numerators calculated for all optional next nodes, which is used for probability normalization. More specifically, the state transition probability for: ; in, for Time Path The pheromone concentration on To provide comprehensive inspiration information, For couriers For Node Proficiency in the area For nodes The urgency of the task; 、 、 、 is the preset non-negative weight parameter; the sum of the denominator part is for the courier All accessible next nodes In another embodiment, before calculating the transition probability, factors such as pheromone, heuristic information, proficiency, and urgency are normalized, for example, scaled to the interval [0, 1].

[0034] When choosing a path, we balance the influence of historical experience, immediate cost, personnel capabilities and task priority, and consider various realistic factors in the intelligent decision-making process. For example, even if a path is short, that is, the comprehensive heuristic information The value is high, but if the congestion is frequent in history, the pheromone The value is low, or the courier is not familiar with the area. The value is low, or the task is not urgent. If the value is low, the probability of being selected may be lower than that of another path that is slightly farther away but has balanced performance in all aspects. For example, suppose the courier is at node , you can choose to go to the node or . Go The distance is short but the courier is unskilled and the task is not urgent; go to The route is long but the courier is very familiar with it and the task is urgent. The pheromone is 1.2, the inspiration information is 2.0, the proficiency is 0.8, and the urgency is 1.0; the path =Pheromone is 1.0, Heuristic is 1.5, Proficiency is 1.5, and Urgency is 2.0. If the weights are set to equal, go to The numerator product is 1.92, and going to The numerator product is 4.5. The result will be given to The path with higher transition probability, such as Figure 2 As shown, more reasonable delivery decisions can be made. 、 、 、 are all scalar values.

[0035] After determining the transfer probability, the roulette wheel selection method is used to select the next node for the courier based on the calculated probability.

[0036] S3, after each iteration is completed, the spatiotemporal volatility factor is calculated based on the historical traffic efficiency and real-time congestion index, and the global pheromone is updated with the spatiotemporal volatility factor; based on the path quality of the current iterative optimal solution, the optimal path is pheromone enhanced; the path quality is a scalar value that comprehensively evaluates the total path duration, time window satisfaction and comprehensive path cost.

[0037] Specifically, the pheromone on all paths is globally updated, and the spatiotemporal volatility factor According to road section The pheromone decay factor is calculated based on the historical traffic efficiency and real-time congestion index. For example, the more congested or the lower the historical efficiency of the road section, the larger the spatiotemporal volatility factor value, which makes the pheromone on the old or inefficient path decay faster.

[0038] Calculate the quality score of the optimal path found in the current iteration The score is calculated by a utility function whose input includes the total duration of the path, the degree of satisfaction of all task time windows (e.g., the sum of penalties for not meeting the time windows), and the sum of the comprehensive costs of all nodes on the path. In one embodiment, only the edges on the optimal path are Increase pheromones, increase the amount and quality score In direct proportion, the path to high-quality solutions receives greater positive feedback.

[0039] S4, repeat the path construction and pheromone update steps until the preset termination condition is met, and output the optimal pickup path sequence.

[0040] In one embodiment, the termination condition is set to one or a combination of the following: reaching a preset maximum number of iterations, such as 500; the global optimal solution has not been improved for a certain number of consecutive generations, such as 50; or the total run time exceeds a preset threshold, such as 30 seconds. Once the termination condition is met, the global optimal solution recorded throughout the entire run is output. This solution is an ordered list of client node IDs, representing the determined optimal pickup path.

[0041] In an optional embodiment, the calculation process of the comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The quantified value of parking difficulty in the area and the estimated waiting time, where parking difficulty is quantified as a value between 0 and 1. The greater the parking difficulty, the greater the quantified value of parking difficulty; The quantified value of parking difficulty is multiplied by the expected waiting time and summed with the predicted travel time to obtain the comprehensive path cost.

[0042] The inverse of the comprehensive path cost is used as comprehensive heuristic information.

[0043] The comprehensive heuristic information The calculation formula is: ; in, For nodes To Node The predicted travel time, For nodes A quantitative value of the parking difficulty in the area, For nodes Estimated wait time.

[0044] The estimated travel time takes into account real-time traffic conditions, the parking difficulty quantifies the time and effort that may be spent looking for a parking space at the destination, and the estimated waiting time takes into account possible delays at the customer or pickup point, such as waiting for the elevator, the customer looking for the package, etc. These factors together form a more comprehensive cost assessment model. For example, from the delivery station To customers and The map distance is close. But In old residential areas with scarce parking spaces, the parking difficulty coefficient may be as high as 4.8, with an estimated waiting time of 5 minutes; Located in an office building with a dedicated unloading area, the parking difficulty coefficient is 1.0, and there is no waiting time. Assuming the driving time is 15 minutes, The calculated comprehensive path cost value will be significantly higher than After taking the reciprocal, The comprehensive heuristic information value will be larger and more attractive in path selection, thereby guiding the system to plan a route with higher actual execution efficiency.

[0045] In other embodiments, a weighted sum may be performed on the product of the predicted travel time and the quantified value of parking difficulty and the estimated waiting time.

[0046] In other embodiments, the node The difficulty of parking in the area can be measured using nodes The parking space search time in the area is characterized. At this time, the calculation process of the comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The time it takes to find a parking space in your area and the estimated waiting time; The predicted travel time, parking space search time and estimated waiting time are summed to obtain the comprehensive path cost; The inverse of the comprehensive path cost is used as comprehensive heuristic information.

[0047] In an optional embodiment, the size of the spatiotemporal volatility factor is positively correlated with the real-time congestion index and negatively correlated with the historical traffic efficiency.

[0048] To adapt to real-time changes in road conditions, the system intelligently manages the pheromone memory lifespan by linking spatiotemporal volatility factors with traffic data. When a road suddenly becomes congested, the spatiotemporal volatility factor is increased to accelerate the removal of pheromones accumulated along that route, preventing subsequent route planning from relying on outdated, reliable experience. For example, during the morning rush hour at 9:00 AM, if the real-time congestion index of urban arterial road A reaches 0.9, a higher spatiotemporal volatility factor, such as 0.5, is set to rapidly decay the previously accumulated pheromones. At 3:00 PM, however, the road's historical efficiency index is 0.95, indicating very smooth traffic. In this case, an extremely low spatiotemporal volatility factor, such as 0.05, is set. This effectively preserves the high-quality path information formed on that route overnight, ensuring its continued use in subsequent planning. This ensures precise synchronization of the pheromone update rate with real-world road conditions.

[0049] In one embodiment, the spatiotemporal volatility factor The calculation formula is: ; in, is the real-time congestion index, The higher the real-time congestion index is, the greater the space-time volatility factor is; the higher the historical traffic efficiency is, the smaller the space-time volatility factor is.

[0050] In an optional embodiment, the pheromone enhancement of the optimal path includes: Calculate the path quality of the current iterative optimal solution, where the path quality is a function value that is proportional to the time window satisfaction and inversely proportional to the total path duration and the comprehensive path cost; An additional pheromone increment proportional to the quality of the path is applied to the edges on the optimal path.

[0051] A good path must not only be short and cost-effective, but more importantly, it must meet customers' time requirements. By directly linking pheromone increments to path quality, paths that excel across multiple dimensions receive stronger positive feedback, effectively guiding the algorithm toward the global optimal solution. For example, in one iteration, two paths were found. Path one, with a total duration of 100 minutes and a cost of 80, met the time windows of all ten customers, achieving a 100% time window satisfaction rating. Path two, with a total duration of 95 minutes and a cost of 75, missed the time window of one customer, resulting in a 90% time window satisfaction rating. Although path one is slightly longer, its perfect time window satisfaction likely results in a higher path quality score. Segments along path one will be rewarded with more pheromone rewards, encouraging subsequent iterations to explore and utilize this higher-quality path.

[0052] In a second aspect of the present invention, a pickup optimization system for smart logistics is provided, comprising the following modules: The data acquisition module is used to obtain a set of pending pickup tasks, courier information, and real-time road network data; the pickup task set includes the location, customer time window, and task urgency of each task; the courier information includes proficiency in each area; and the real-time road network data includes road network topology, predicted travel time between nodes, real-time congestion index, historical traffic efficiency, regional parking difficulty, and estimated waiting time at the target pickup point; The transfer probability calculation module is used to calculate the courier's probability of moving from the current node to the next node in the path construction of each iteration of the ant colony system based on the path pheromone concentration, comprehensive heuristic information, courier's regional proficiency and task urgency. Go to the next unvisited node The transition probability of the comprehensive heuristic information is the inverse of the comprehensive path cost, which is a scalar value that comprehensively evaluates the predicted travel time between nodes, the difficulty of parking in the area, and the estimated waiting time of the target pickup point; An update module is configured to calculate a spatiotemporal volatility factor based on historical traffic efficiency and real-time congestion index after each iteration, and to update the global pheromone using the spatiotemporal volatility factor; perform pheromone enhancement on the optimal path based on the path quality of the current iterative optimal solution; the path quality is a scalar value that comprehensively evaluates the total path duration, time window satisfaction, and comprehensive path cost; The path output module is used to repeat the path construction and pheromone update steps until the preset termination conditions are met and output the optimal pickup path sequence.

[0053] Optionally, calculate the courier from the current node Go to the next unvisited node The transition probability , calculated using the following formula: ; in, for Time Path The pheromone concentration on To provide comprehensive inspiration information, For couriers For Node Proficiency in the area For nodes The urgency of the task; 、 、 、 is the preset non-negative weight parameter; the sum of the denominator part is for the courier All accessible next nodes Perform calculations.

[0054] Optionally, the calculation process of the comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The quantified value of parking difficulty in the area and the estimated waiting time, where the quantified value of parking difficulty is proportional to the parking difficulty; The quantified value of parking difficulty is multiplied by the estimated waiting time, and then summed with the predicted travel time to obtain the comprehensive path cost; The inverse of the comprehensive path cost is used as the comprehensive heuristic information.

[0055] Optionally, the size of the spatiotemporal volatility factor is positively correlated with the real-time congestion index and negatively correlated with the historical traffic efficiency.

[0056] Optionally, the pheromone enhancement of the optimal path comprises the following steps: Calculate the path quality of the current iterative optimal solution, where the path quality is a function value that is proportional to the time window satisfaction and inversely proportional to the total path duration and the comprehensive path cost; An additional pheromone increment proportional to the quality of the path is applied to the edges on the optimal path.

[0057] Optionally, the urgency of the task is inversely proportional to the time limit required by the customer to complete the task; and the regional proficiency of the courier is positively correlated with the courier's familiarity with the area where the target node is located.

[0058] The above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in accordance with the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0059] The steps of the methods or algorithms described in the embodiments of the present invention may be directly embedded in hardware, software components executed by a processor, or a combination of the two. The software components may be stored in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. For example, the storage medium may be connected to the processor so that the processor can read information from the storage medium and write information to the storage medium. Alternatively, the storage medium may be integrated into the processor. The processor and storage medium may be provided in an ASIC.

[0060] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0061] In addition, in the description of this specification, “a plurality of” means at least two, for example, two, three or more, etc., unless otherwise clearly and specifically defined.

Claims

1. A pickup optimization method for smart logistics, characterized in that: The following steps are involved: S1, obtains the set of pending pickup tasks, courier information and real-time road network data; the pickup task set includes the location of each task, customer time window and task urgency; the courier information includes the proficiency in each area; the real-time road network data includes the road network topology, predicted travel time between nodes, real-time congestion index, historical traffic efficiency, regional parking difficulty and estimated waiting time of the target pickup point; S2, in each iteration of the path construction of the ant colony system, the courier's journey from the current node is calculated based on the path pheromone concentration, comprehensive heuristic information, courier regional proficiency and task urgency. Go to the next unvisited node The transition probability is denoted by the comprehensive heuristic information, which is the inverse of the comprehensive path cost. The comprehensive path cost is a scalar value that comprehensively evaluates the predicted travel time between nodes, the difficulty of parking in the area, and the estimated waiting time at the target pickup point. After each iteration, S3 calculates the spatiotemporal volatility factor based on the historical traffic efficiency and the real-time congestion index, and uses the spatiotemporal volatility factor to update the global pheromone. Based on the path quality of the current iterative optimal solution, the optimal path is enhanced with pheromones. Path quality is a scalar value that comprehensively evaluates the total path duration, time window satisfaction, and comprehensive path cost. In step S4, the path construction and pheromone update steps are repeated until the preset termination condition is met, and the optimal pickup path sequence is output.

2. The method for optimizing pickup for smart logistics according to claim 1, characterized in that: Calculate the courier from the current node Go to the next unvisited node The transition probability , calculated using the following formula: ; in, for Time Path The pheromone concentration on To provide comprehensive inspiration information, For couriers For Node Proficiency in the area For nodes The urgency of the task; 、 、 、 is the preset non-negative weight parameter; the sum of the denominator part is for the courier All accessible next nodes Perform calculations.

3. The method for optimizing pickup for smart logistics according to claim 1, characterized in that: The calculation process of comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The quantified value of parking difficulty in the area and the estimated waiting time, where the quantified value of parking difficulty is proportional to the parking difficulty; The quantified value of parking difficulty is multiplied by the estimated waiting time, and then summed with the predicted travel time to obtain the comprehensive path cost; The inverse of the comprehensive path cost is used as the comprehensive heuristic information.

4. The method for optimizing pickup for smart logistics according to claim 1, characterized in that: The size of the spatiotemporal volatility factor is positively correlated with the real-time congestion index and negatively correlated with the historical traffic efficiency.

5. The method for optimizing pickup for smart logistics according to claim 1, characterized in that: The pheromone enhancement of the optimal path includes the following steps: Calculate the path quality of the current iterative optimal solution. The path quality is a function value that is proportional to the time window satisfaction and inversely proportional to the total path duration and the comprehensive path cost. Apply an additional pheromone increment proportional to the path quality to the edges on the optimal path.

6. The method for optimizing pickup for smart logistics according to any one of claims 1 to 5, characterized in that: The urgency of the task is inversely proportional to the time limit required by the customer to complete the task; the courier's proficiency in each area is directly proportional to the courier's familiarity with the area where the target node is located.

7. A pickup optimization system for smart logistics, characterized by: include: The data collection module is used to obtain a set of pending pickup tasks, courier information, and real-time road network data. The pickup task set includes the location, customer time window, and task urgency of each task. The courier information includes proficiency in each area. The real-time road network data includes the road network topology, predicted travel time between nodes, real-time congestion index, historical traffic efficiency, regional parking difficulty, and estimated waiting time at the target pickup point. The transfer probability calculation module is used to calculate the courier's probability of moving from the current node to the next node in the path construction of each iteration of the ant colony system based on the path pheromone concentration, comprehensive heuristic information, courier's regional proficiency and task urgency. Go to the next unvisited node The transition probability is , where the comprehensive heuristic information is the inverse of the comprehensive path cost, which is a scalar value that comprehensively evaluates the predicted travel time between nodes, the difficulty of regional parking, and the estimated waiting time of the target pickup point. The update module is used to calculate the spatiotemporal volatility factor based on the historical traffic efficiency and real-time congestion index after each iteration, and use the spatiotemporal volatility factor to update the global pheromone. Based on the path quality of the current iterative optimal solution, the optimal path is pheromone enhanced. The path quality is a scalar value that comprehensively evaluates the total path duration, time window satisfaction, and comprehensive path cost. The path output module is used to repeat the path construction and pheromone update steps until the preset termination condition is met and output the optimal pickup path sequence.

8. The pickup optimization system for smart logistics according to claim 7, characterized in that: Calculate the courier from the current node Go to the next unvisited node The transition probability , calculated using the following formula: ; in, for Time Path The pheromone concentration on To provide comprehensive inspiration information, For couriers For Node Proficiency in the area For nodes The urgency of the task; 、 、 、 is the preset non-negative weight parameter; the sum of the denominator part is for the courier All accessible next nodes Perform calculations.

9. The pickup optimization system for smart logistics according to claim 7, characterized in that: The calculation process of the comprehensive heuristic information is: Compute slave nodes To Node The predicted travel time and the node The quantified value of parking difficulty in the area and the estimated waiting time, where the quantified value of parking difficulty is proportional to the parking difficulty; The quantified value of parking difficulty is multiplied by the estimated waiting time, and then summed with the predicted travel time to obtain the comprehensive path cost; The inverse of the comprehensive path cost is used as comprehensive heuristic information.

10. The pickup optimization system for smart logistics according to claim 7, 8 or 9, characterized in that: The size of the spatiotemporal volatility factor is positively correlated with the real-time congestion index and negatively correlated with the historical traffic efficiency.

Images