A method for planning transportation routes for logistics vehicles

Abstract

This invention discloses a transportation route planning method for logistics vehicles, comprising the following steps: S1, multi-source real-time data access and standardization; S2, task modeling and constraint construction; S3, candidate route search; S4, multi-objective evaluation and route output; S5, target route issuance and adjustment. The purpose of this invention is to provide a transportation route planning method for logistics vehicles, aiming to utilize multi-source data to construct hard and soft constraints as the data foundation for route planning, and to evaluate the optimal route and replan the route based on abnormal events according to multiple objectives, thereby overcoming the problems described in the background art.

Description

Technical Field

[0001] This invention relates to the field of transportation technology, and more specifically to a method for planning transportation routes for logistics vehicles. Background Technology

[0002] In multi-node logistics transportation scenarios, transportation route planning usually needs to simultaneously meet multiple constraints such as time windows, road physical restrictions (height / weight / restricted hours), vehicle range and load capacity, and make real-time adjustments in abnormal situations such as road congestion, road closures, order changes, and vehicle malfunctions.

[0003] In existing technologies, logistics vehicle transportation route planning typically relies on static maps or single objectives (such as shortest distance), failing to effectively integrate real-time road conditions and vehicle dynamics. This results in significant deviations between planned routes and actual traffic. Taking tobacco industry transportation as an example, there are also industry-specific constraints such as unloading windows, loading and unloading sequences, consolidation along routes, and restrictions on park access. When these constraints are superimposed on dynamic road conditions, traditional route planning schemes struggle to simultaneously consider distance, cost, carbon emissions, and on-time performance, impacting transportation service quality. Therefore, a route planning method is needed that can integrate multi-source data, embed hard constraint processing during the search phase, support multi-objective collaboration, and possess event-driven replanning and system linkage capabilities. Summary of the Invention

[0004] The purpose of this invention is to provide a transportation route planning method for logistics vehicles, which aims to utilize multi-source data to construct hard and soft constraints as the data basis for route planning, and to evaluate the optimal route and replan the route based on multiple objectives and abnormal events, thereby overcoming the problems described in the background art.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] A method for planning transportation routes for logistics vehicles, comprising the following steps:

[0007] S1. Multi-source real-time data access and standardization: The vehicle terminal collects the positioning data and vehicle status data of logistics vehicles and uploads them to the route optimization server. The route optimization server obtains road condition data from the road condition service interface, road attribute data from the road attribute database, order constraint data from the order system, and park passage constraint data from the electronic fence system. The collected data is then filtered for anomalies, filled in missing data, and standardized to generate a route optimization dataset.

[0008] S2. Task modeling and constraint construction: The route optimization server constructs a delivery task set based on order constraint data, constructs a road hard constraint set based on the road attribute data and park access constraint data, and constructs a vehicle hard constraint set and a soft constraint set based on vehicle status data.

[0009] S3. Candidate route search: The route optimization server performs a candidate route search, introducing constraint satisfaction during the state transition from node i to node j. Adjust the state transition probability, where when candidate edge (i,j) violates either of the aforementioned road hard constraints or vehicle hard constraints, let =0, when candidate edge (i,j) satisfies all the road hard constraints and vehicle hard constraints. >0, to generate a set of candidate routes that satisfy the hard constraints;

[0010] S4. Multi-objective evaluation and route output: The route optimization server performs multi-objective evaluation on the candidate route set to determine the target route, and determines the weight of each objective based on the transportation scenario, and outputs the target route and the estimated arrival time of each node.

[0011] S5. Target route distribution and adjustment: The route optimization server distributes the target route to the vehicle terminal for navigation execution. When a preset abnormal event is detected, replanning is triggered. During replanning, the vehicle's current position is used as the starting point, and the completed nodes and remaining constraints are inherited to generate an updated route. The updated route is distributed to the vehicle terminal, and the electronic fence system is authorized to update the channel and the updated estimated arrival time is pushed to the customer system.

[0012] Furthermore, in step S1, the vehicle status data includes at least one or more of the following: remaining driving range, battery level, fuel level, and load capacity.

[0013] The anomaly filtering specifically includes:

[0014] When the offline duration of the vehicle terminal positioning signal exceeds the first threshold, the corresponding positioning data is removed.

[0015] When the number of consecutive missing road condition data exceeds the second threshold, historical road condition data is used to complete the data.

[0016] The standardization process specifically includes:

[0017] Convert unloading windows into timestamp intervals, and convert road height restrictions, weight restrictions, restricted hours, and park access restrictions into feasibility assessment rules;

[0018] The missing information completion specifically includes: completing the missing road conditions using historical road conditions and vehicle energy consumption models from the same period.

[0019] Furthermore, in step S2, the set of hard road constraints includes height restrictions, unloading window constraints, weight restrictions, restricted traffic periods, temporary traffic control constraints, and park access constraints.

[0020] The set of vehicle hard constraints includes vehicle range constraints;

[0021] The soft constraints include loading and unloading sequence constraints, route grouping constraints, peak avoidance constraints, and priority preference constraints.

[0022] Furthermore, in step S3, the correction of the state transition probability specifically includes:

[0023] S301. Obtain the heuristic factor, which can be derived using the following formula:

[0024]

[0025] In the formula, As a heuristic factor, This refers to dynamic driving time;

[0026] S302. The constraint satisfaction degree is defined by the following formula:

[0027]

[0028] In the formula, This is the gain coefficient. For soft-constraint preferences, the value range can be set to... , To constrain satisfaction;

[0029] S303. Calculate the state transition probability, the formula is as follows:

[0030]

[0031] In the formula, For the candidate adjacency set, For parameters, As a heuristic factor, To constrain satisfaction, The intensity of the pheromone.

[0032] Furthermore, in step S4, the route optimization server performs a multi-objective evaluation on the candidate route set to determine the target route. The multi-objectives include distance, transportation cost, carbon emissions, and on-time delivery rate costs. The sum of the weights of each objective is 1. A comprehensive multi-objective cost is constructed based on the weights of each objective, using the following formula:

[0033] ;

[0034] In the formula, This refers to the route distance; For transportation costs, For the cost of carbon emissions, The cost of unpunctuality.

[0035] Furthermore, in step S5, the preset abnormal events include sudden changes in road conditions, abnormal vehicle status, order changes, changes in unloading windows, temporary closure of park access, and abnormal deviation.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] This invention introduces multi-source data, including vehicle positioning data, vehicle status data, road condition data, road attribute data, order constraint data, and park access constraint data, to provide a data foundation for route planning. Through the fusion and standardization of multi-source data, the adaptability of travel time estimation and route selection to dynamic road conditions is improved.

[0038] Constructing hard and soft constraints using multi-source data, and then applying hard constraints through... By embedding the state transition probability in the search phase, the proportion of feasible solutions is significantly improved and invalid iterations are reduced, thereby generating feasible candidate road segments;

[0039] This invention introduces a multi-objective dynamic weighting system that takes into account factors such as distance, cost, carbon emissions, and on-time performance, thereby reducing transportation costs while meeting order requirements.

[0040] In this invention, when an abnormal event is detected, the route is adjusted starting from the vehicle's current position and inheriting completed nodes and remaining constraints. This improves the speed of abnormal response and the feasibility of the route, reducing manual intervention. Attached Figure Description

[0041] Figure 1 This is a flowchart of the transportation route planning method described in this invention.

[0042] Figure 2 This is a schematic diagram of the architecture of a transportation route planning method. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0044] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0045] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0046] Example 1

[0047] refer to Figure 1-2 As shown, the present invention provides a method for planning transportation routes for logistics vehicles, including the following steps:

[0048] S1. Multi-source real-time data access and standardization: The vehicle terminal collects the positioning data and vehicle status data of logistics vehicles and uploads them to the route optimization server. The route optimization server obtains road condition data from the road condition service interface, road attribute data from the road attribute database, order constraint data from the order system, and park passage constraint data from the electronic fence system. The collected data is then filtered for anomalies, filled in missing data, and standardized to generate a route optimization dataset.

[0049] In this step, the vehicle status data includes at least one or more of the following: remaining driving range, battery level, fuel level, and load capacity. Specifically, the vehicle terminal obtains the vehicle's real-time location (WGS84 latitude and longitude) and vehicle status (remaining battery / fuel level, current speed, vehicle type, and load capacity, etc.) through BeiDou dual-mode positioning. Its sampling frequency is once every 10 seconds, with a dynamic accuracy of ±5cm. The route optimization server uses an NVIDIA A10 processor with a computing power of 25 TOPS, deploys an improved ACO algorithm, and interfaces with the vehicle terminal, Gaode Map's real-time traffic API, the logistics management system, and the electronic fence system. This allows it to access real-time traffic conditions from Gaode Map, including congestion indices for each road segment, estimated travel time, unexpected events, and road height and width restrictions, as well as road attribute data. It also interfaces with the logistics management system to obtain order constraint data, such as multi-node delivery requirements, unloading windows at each node, and product specification loading / unloading sequences. Finally, it interfaces with the electronic fence system to obtain channel constraints for logistics parks / commercial companies, such as only lane 1 allowing heavy vehicles to enter and exit the logistics park.

[0050] The missing information completion specifically includes: completing the missing road conditions using historical road conditions and vehicle energy consumption models from the same period;

[0051] The anomaly filtering specifically includes:

[0052] When the offline time of the vehicle terminal positioning signal exceeds the first threshold, the corresponding positioning data is removed, such as removing Beidou signal offline for more than 15 seconds;

[0053] When the number of consecutive missing road condition data exceeds the second threshold, historical road condition data is used to fill in the missing data. For example, if three consecutive API call failures result in consecutive missing road condition data, the missing values ​​are filled in using "historical road condition data from the same period + vehicle range model".

[0054] The standardization process specifically includes:

[0055] The unloading windows were converted into timestamp intervals, and road height restrictions, weight restrictions, prohibited hours, and park access restrictions were converted into feasibility assessment rules. In actual processing, all location data was converted to WGS84 latitude and longitude, and road segment lengths were uniformly converted to kilometers. Unloading windows were converted into timestamp intervals (e.g., Hangzhou Tobacco 9:00-17:00 converted to 1717202039000-1717230839000), and loading / unloading order, height and width restrictions, and other constraints were converted into Boolean rules (e.g., "Slim cigarettes unload first → Rule 1: Node 1 (Hangzhou) loading / unloading order = slim cigarettes → thick cigarettes").

[0056] S2. Task modeling and constraint construction: The route optimization server constructs a delivery task set based on order constraint data, constructs a road hard constraint set based on the road attribute data and park access constraint data, and constructs a vehicle hard constraint set and a soft constraint set based on vehicle status data.

[0057] In this step, the set of hard road constraints includes height restrictions, unloading window constraints, weight restrictions, restricted hours constraints, temporary traffic control constraints, and park access constraints.

[0058] The set of vehicle hard constraints includes vehicle range constraints;

[0059] The soft constraints include loading and unloading sequence constraints, route grouping constraints, peak avoidance constraints, and priority preference constraints.

[0060] Regarding the above hard and soft constraints, it is particularly important to note the following constraints: vehicle range constraint: total route mileage ≤ maximum mileage supported by remaining battery / fuel; unloading window constraint: vehicle arrival time must be within the unloading window; height restriction constraint: route height restriction ≥ vehicle height; park access constraint: designated access channels must be used to enter and exit the logistics park; loading and unloading sequence constraint: multi-node delivery must be planned in a certain order (e.g., loading and unloading at the small cigarette node first, then the large cigarette node); route consolidation constraint: orders in the same direction must be consolidated (e.g., Kunming → Hangzhou can be consolidated into Kunming → Shanghai); peak avoidance constraint: avoid peak hours in the unloading area; priority preference constraint: high-priority order routes must prioritize on-time delivery.

[0061] An exemplary scenario using this hard constraint is as follows: the Kunming to Hangzhou route needs to avoid the "G320 National Highway section" with a height restriction of 4.0 meters and choose the "Shanghai-Kunming Expressway" with a height restriction of 4.5 meters; the unloading of Hangzhou Tobacco needs to arrive between 9:00 and 17:00;

[0062] An exemplary scenario using this soft constraint is as follows: for a Kunming → Hangzhou → Shanghai consolidation route, it is necessary to ensure that Hangzhou is reached first and Shanghai is reached later; and to avoid the peak unloading time of Hangzhou tobacco from 10:00 to 12:00.

[0063] In application, hard constraints take precedence over soft constraints. A two-tiered constraint system of "hard constraints + soft constraints" ensures both compliance and optimization of the route.

[0064] S3. Candidate route search: The route optimization server performs a candidate route search, introducing constraint satisfaction during the state transition from node i to node j. Adjust the state transition probability, where when candidate edge (i,j) violates either of the aforementioned road hard constraints or vehicle hard constraints, let =0, when candidate edge (i,j) satisfies all the road hard constraints and vehicle hard constraints. >0, to generate a set of candidate routes that satisfy the hard constraints;

[0065] Specifically, the corrected state transition probability includes:

[0066] S301. Obtain the heuristic factor, which can be derived using the following formula:

[0067]

[0068] In the formula, As a heuristic factor, This refers to dynamic driving time;

[0069] S302. The constraint satisfaction degree is defined by the following formula:

[0070]

[0071] In the formula, This is the gain coefficient. For soft-constraint preferences, the value range can be set to... , To constrain the satisfaction degree, thus making The value is 0 or ;

[0072] S303. Calculate the state transition probability, the formula is as follows:

[0073]

[0074] In the formula, For the candidate adjacency set, For parameters, As a heuristic factor, To constrain satisfaction, To determine pheromone strength, pheromone updates can employ feedback related to the multi-objective comprehensive cost. For example, routes with better comprehensive costs can be enhanced by updating coefficients, thereby accelerating convergence. This can be achieved by allowing candidate edges that violate hard constraints to... This allows it to be directly excluded during the search phase, improving the efficiency of generating feasible paths.

[0075] S4. Multi-objective evaluation and route output: The route optimization server performs multi-objective evaluation on the candidate route set to determine the target route, and determines the weight of each objective based on the transportation scenario, and outputs the target route and the estimated arrival time of each node.

[0076] The multi-objectives include distance, transportation cost, carbon emissions, and the cost of untimely arrival. The sum of the weights of each objective is 1. The comprehensive cost of the multi-objectives is constructed based on the weights of each objective, and its formula is as follows:

[0077] ;

[0078] In the formula, This refers to the route distance; For transportation costs, For the cost of carbon emissions, The cost of unpunctuality.

[0079] It should be noted that the weights of each objective are not fixed values. In practical applications, the weights of each objective can be determined according to the transportation scenario, as shown in the table below:

[0080] Scene type Distance weight ( ) Cost weight ( ) Carbon emission weights ( ) On-time rate weight ( ) Weight constraints ( + + + =1) Typical scenarios 0.2 0.3 0.2 0.3 Balancing multiple objectives, suitable for daily transportation Emergency Scenarios 0.1 0.2 0.1 0.6 Prioritizing on-time delivery rate applies to P3 orders. Green Scene 0.15 0.25 0.4 0.2 Prioritizing carbon emission reduction applies to environmental assessment periods.

[0081] S5. Target route distribution and adjustment: The route optimization server distributes the target route to the vehicle terminal for navigation execution. When a preset abnormal event is detected, replanning is triggered. During replanning, the vehicle's current position is used as the starting point, and the completed nodes and remaining constraints are inherited to generate an updated route. The updated route is distributed to the vehicle terminal, and the electronic fence system is authorized to update the channel and the updated estimated arrival time is pushed to the customer system.

[0082] In the steps, the preset abnormal events include sudden changes in road conditions, abnormal vehicle status, order changes, changes in unloading windows, temporary closure of park access, and abnormal deviation.

[0083] In practical applications, when an abnormal event is detected, the following operations are performed:

[0084] 1) Use the road network node corresponding to the vehicle's current location as the new starting point;

[0085] 2) Lock the completed nodes and inherit the set of remaining nodes and constraints;

[0086] 3) Perform rapid replanning (e.g., reduce the number of iterations and candidate size, and output a new route within a preset time).

[0087] 4) Send the new route to the vehicle terminal;

[0088] 5) Update the access authorization for the electronic fence system;

[0089] 6) Push updated ETA to the customer's system and synchronize reservation information.

[0090] Example 2

[0091] Using the transportation route planning method for logistics vehicles described in this invention, the following detailed implementation process is illustrated in the multi-node transportation scenario of Hongyun Honghe Group's "Kunming Plant → Hangzhou Tobacco → Shanghai Tobacco" route (26.6-ton new energy unmanned truck V001, P1 peak season order, Hangzhou unloading window 9:00-17:00, Shanghai unloading window 14:00-20:00):

[0092] 1. Implement environmental preparation

[0093] Hardware environment: Beidou dual-mode positioning terminal (equipped in V001), GPU server (NVIDIA A10, computing power 25TOPS, deployed with improved ACO algorithm), and scheduling center terminal (displaying route planning results).

[0094] Software environment: The backend uses the Spring Boot framework, the map service is integrated with the Gaode Map API (Enterprise Edition), the algorithm is implemented using Python (numpy, scipy libraries), and the data storage uses MySQL + Redis (caching real-time traffic conditions).

[0095] Data interfaces: Connects to BeiDou positioning platform, Gaode Map real-time traffic API, logistics management system (order constraints), and electronic fence system (park access constraints).

[0096] 2. Transportation route planning methods:

[0097] Step 1: Data Access

[0098] Vehicle data: V001 real-time location (Kunming plant, 102.7386°E, 25.1161°N), remaining battery power supports 2200km range, load capacity 22 tons (10 tons for slim cigarettes, 12 tons for thick cigarettes).

[0099] Traffic data: Gaode API returned "Kunming → Hangzhou Shanghai-Kunming Expressway congestion index 4.2 (severe congestion), estimated delay 1.5 hours; Hangrui Expressway is smooth, estimated travel time 21 hours";

[0100] Order restrictions: Hangzhou unloading window 9:00-17:00, Shanghai 14:00-20:00; loading and unloading order: "Slim cigarettes unloaded first in Hangzhou, larger cigarettes unloaded later in Shanghai".

[0101] Electronic fence: Hangzhou Tobacco only allows heavy vehicles to enter the "East Unloading Channel". Shanghai Tobacco's unloading area is in peak hours from 14:00 to 16:00, so please avoid it.

[0102] Data processing: Remove data from congested sections of the Shanghai-Kunming Expressway, convert unloading windows into timestamps, and confirm that the V001's range meets the requirements of the Hangzhou-Ruili Expressway route (2000km≤2200km).

[0103] Step Two

[0104] Establish hard constraints: a) Route height limit ≥ 4.2 meters; b) Arrive in Hangzhou before 17:00 and in Shanghai before 20:00; c) Use Hangzhou East unloading channel.

[0105] Establish soft constraints: a) Unload slim cigarettes first in Hangzhou, and then unload coarse cigarettes in Shanghai; b) Avoid the unloading peak in Shanghai from 14:00 to 16:00, and expect to arrive in Shanghai after 16:00.

[0106] Step 3

[0107] Introducing constraint satisfaction into the ACO algorithm The state transition probabilities are adjusted and multiple iterations are performed to output a set of candidate routes.

[0108] Step Four

[0109] This is configured as a regular transportation scenario, thereby assigning multi-objective weights to the candidate route set. =0.2, =0.3, =0.2, =0.3; thus outputting the optimal target route: Kunming plant area → Hangrui Expressway → Changshen Expressway → Hangzhou Tobacco (East Unloading Channel) → Hukun Expressway → Shanghai Tobacco (Backup Channel). Multi-objective results: total distance 1980km, cost 4200 yuan, carbon emissions 4200kg, expected arrival in Hangzhou at 15:30 (within the window), arrival in Shanghai at 16:30 (avoiding peak hours), and the optimal target route and multi-objective results are synchronized to the customer terminal.

[0110] Step 5

[0111] The optimal target route is sent to the vehicle terminal for navigation of vehicle V001. When V001 reaches the Zhejiang section of the Hangrui Expressway, the Gaode API pushes "Construction ahead 5km, detour required, estimated delay 0.5 hours". At this time, ACO is invoked, iterates 20 times and outputs a new route (the distance increases by 15km after the detour, estimated arrival time in Hangzhou is 16:00), which is still within the window; the electronic fence updates the access permissions of the detour section at the same time, and the Hangzhou Tobacco terminal receives the new estimated arrival time of 16:00.

[0112] Although the invention has been described herein with reference to several illustrative embodiments, it should be understood that many other modifications and implementations can be devised by those skilled in the art, which will fall within the scope and spirit of the principles disclosed herein. More specifically, various modifications and improvements can be made to the components or layout of the subject matter arrangement within the scope of the disclosure, drawings, and claims. Besides modifications and improvements to the components or layout, other uses will be apparent to those skilled in the art.

Claims

1. A method for planning transportation routes for logistics vehicles, characterized in that, Includes the following steps: S1. Multi-source real-time data access and standardization: The vehicle terminal collects the positioning data and vehicle status data of logistics vehicles and uploads them to the route optimization server. The route optimization server obtains road condition data from the road condition service interface, road attribute data from the road attribute database, order constraint data from the order system, and park passage constraint data from the electronic fence system. The collected data is then filtered for anomalies, filled in missing data, and standardized to generate a route optimization dataset. S2. Task modeling and constraint construction: The route optimization server constructs a delivery task set based on order constraint data, constructs a road hard constraint set based on the road attribute data and park access constraint data, and constructs a vehicle hard constraint set and a soft constraint set based on vehicle status data. S3. Candidate route search: The route optimization server performs a candidate route search, introducing constraint satisfaction during the state transition from node i to node j. Adjust the state transition probability, where when candidate edge (i,j) violates either of the aforementioned road hard constraints or vehicle hard constraints, let =0, when candidate edge (i,j) satisfies all the road hard constraints and vehicle hard constraints. >0, to generate a set of candidate routes that satisfy the hard constraints; S4. Multi-objective evaluation and route output: The route optimization server performs multi-objective evaluation on the candidate route set to determine the target route, and determines the weight of each objective based on the transportation scenario, and outputs the target route and the estimated arrival time of each node. S5. Target route distribution and adjustment: The route optimization server distributes the target route to the vehicle terminal for navigation execution. When a preset abnormal event is detected, replanning is triggered. During replanning, the vehicle's current position is used as the starting point, and the completed nodes and remaining constraints are inherited to generate an updated route. The updated route is distributed to the vehicle terminal, and the electronic fence system is authorized to update the channel and the updated estimated arrival time is pushed to the customer system.

2. The method for planning transportation routes for logistics vehicles according to claim 1, characterized in that: In step S1, the vehicle status data includes at least one or more of the following: remaining driving range, battery level, fuel level, and load capacity. The anomaly filtering specifically includes: When the offline duration of the vehicle terminal positioning signal exceeds the first threshold, the corresponding positioning data is removed. When the number of consecutive missing road condition data exceeds the second threshold, historical road condition data is used to complete the data. The standardization process specifically includes: Convert unloading windows into timestamp intervals, and convert road height restrictions, weight restrictions, restricted hours, and park access restrictions into feasibility assessment rules; The missing information completion specifically includes: using historical road conditions and vehicle energy consumption models to complete missing road conditions.

3. The method for planning transportation routes for logistics vehicles according to claim 1, characterized in that: In step S2, the set of hard road constraints includes height restrictions, unloading window constraints, weight restrictions, restricted traffic periods, temporary traffic control constraints, and park access constraints. The set of vehicle hard constraints includes vehicle range constraints; The soft constraints include loading and unloading sequence constraints, route grouping constraints, peak avoidance constraints, and priority preference constraints.

4. The method for planning transportation routes for logistics vehicles according to claim 1, characterized in that: In step S3, the correction of the state transition probability specifically includes: S301. Obtain the heuristic factor, which can be derived using the following formula: In the formula, As a heuristic factor, This refers to dynamic driving time; S302. The constraint satisfaction degree is defined by the following formula: In the formula, This is the gain coefficient. For soft-constraint preferences, the value range can be set to... , To constrain satisfaction; S303. Calculate the state transition probability, the formula is as follows: In the formula, For the candidate adjacency set, For parameters, As a heuristic factor, To constrain satisfaction, The intensity of the pheromone.

5. The method for planning transportation routes for logistics vehicles according to claim 1, characterized in that: In step S4, the route optimization server performs a multi-objective evaluation on the candidate route set to determine the target route. The multi-objectives include distance, transportation cost, carbon emissions, and on-time delivery rate costs. The sum of the weights of each objective is 1. A comprehensive multi-objective cost is constructed based on the weights of each objective, using the following formula: ； In the formula, This refers to the route distance; For transportation costs, For the cost of carbon emissions, The cost of unpunctuality.

6. The method for planning transportation routes for logistics vehicles according to claim 1, characterized in that: In step S5, the preset abnormal events include sudden changes in road conditions, abnormal vehicle status, order changes, changes in unloading windows, temporary closure of park access, and abnormal deviation.

Images