A container transportation scheduling method and device, a terminal and a medium

Abstract

The application provides a container transportation scheduling method and device, a terminal and a medium, and relates to the technical field of intelligent logistics scheduling. The method comprises the following steps: obtaining a container transportation request set containing trailer operation information, the trailer operation information containing a first and a second identifier, which respectively indicate whether a tractor with a matching size trailer arrives at a pickup point and whether the trailer is left at a delivery point after delivery; constructing a transportation scheduling planning model with the maximum net income as the target based on the set, solving the ideal transportation scheduling scheme by using a preset heuristic algorithm; simulating the dynamic consumption of the trailer resource based on the initial scheme, determining a shortage request list, and generating a trailer relocation task set according to the shortage request list; constructing a scheduling model with the minimum total cost of trailer relocation as the target based on the task set and solving the scheduling model to obtain a trailer relocation scheduling scheme; and generating a container transportation and trailer resource scheduling scheme based on the two schemes. The application ensures the dynamic balance of the trailer resource and the operability of the scheduling scheme.

Description

Technical Field

[0001] This invention relates to the field of intelligent logistics scheduling technology, and in particular to a container transportation scheduling method, device, terminal and medium. Background Technology

[0002] Currently, optimization of container transportation assumes that tractor-trailers are always paired and that trailers are not allowed to be left at the delivery point. However, in practice, customers often require transportation companies to provide empty trailers at the pickup point or leave fully loaded trailers after delivery for subsequent processing due to loading and unloading equipment or time constraints. This practical need for "trailer separability" makes trailers a scarce resource that dynamically flows between the yard and the customer's location, easily leading to the depletion of yard resources. This makes pre-established static scheduling plans impossible to execute. In other words, while leaving trailers at the delivery point or dispatching new trailers from the trailer yard are important means of solving transportation scheduling problems in container logistics transportation scenarios, leaving trailers at the delivery point may result in insufficient trailers in the trailer yard or increase transportation costs, which is detrimental to the feasibility and efficiency of container hauling operations.

[0003] Moreover, existing technologies mainly focus on tractor route optimization or empty container relocation. Even when the separation of tractor and trailer is involved, they fail to integrate and coordinate the constraints of "trailers being available and available for placement" with the subsequent necessary "dynamic relocation of trailer resources." This results in resource conflicts in the actual implementation of container transport scheduling schemes, or the need for high temporary relocation costs to remedy the situation, lacking robustness and economy in complex actual operating environments. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a container transportation scheduling method, device, terminal and medium, which aims to solve the problems of resource shortage and scheduling infeasibility caused by the ability to leave trailers in the prior art.

[0005] The technical solution adopted by this invention to solve the technical problem is as follows:

[0006] In a first aspect, the present invention discloses a container transportation scheduling method, wherein the method includes:

[0007] Obtain a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether the tractor needs to carry a trailer of the same size to the pickup point, and a second identifier indicating whether the tractor should leave the trailer at the delivery point after the delivery is completed.

[0008] Based on the set of container transport requests, a transport scheduling planning model is constructed with the goal of maximizing net revenue. The ideal transport scheduling scheme is obtained by solving the transport scheduling planning model using a preset heuristic algorithm. The preset heuristic algorithm is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and taboo memory.

[0009] Based on the ideal transportation scheduling scheme, the dynamic consumption of trailer resources during the scheduling execution process is simulated to identify shortage requests where trailer resources are scarce, obtain a shortage request list, and generate a trailer relocation task set based on the shortage request list.

[0010] Based on the set of trailer relocation tasks, a trailer relocation scheduling model is constructed with the goal of minimizing the total cost of trailer relocation, and the trailer relocation scheduling model is solved to obtain a trailer relocation scheduling scheme.

[0011] A container transportation and trailer resource scheduling scheme is generated based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme.

[0012] Optionally, the step of using a preset heuristic algorithm to solve the transportation scheduling planning model to obtain an ideal transportation scheduling scheme includes:

[0013] Construct an initial feasible solution for the transportation scheduling planning model;

[0014] Starting from the initial feasible solution, execute the main iteration loop until the preset iteration termination condition is met. Select the solution with the highest net benefit from all feasible solutions recorded during the iteration process to obtain the ideal transportation scheduling scheme.

[0015] The following sub-steps are executed in each main iteration:

[0016] Randomly select a request from the current solution as a seed request, and calculate the similarity between the seed request and other requests in the current solution;

[0017] Based on the similarity, a first preset number of target requests are determined from other requests, and the target requests are removed from the current solution to obtain a partial solution; the minimum similarity among the first preset number of target requests is not less than the highest similarity among the non-target requests in other requests.

[0018] The target request to be removed is identified as a request to be re-inserted, resulting in multiple requests to be re-inserted;

[0019] Each of the multiple re-insertion requests is sequentially traversed, and the currently traversed re-insertion request is determined as the current re-insertion request;

[0020] Calculate the incremental revenue of the current re-insertion request at different insertion positions in the partial solution;

[0021] The optimal insertion position and a second preset number of suboptimal insertion positions are determined based on the revenue increment; the revenue increment of the current re-insertion request at each of the suboptimal insertion positions is not greater than the revenue increment of the current re-insertion request at the optimal insertion position, but is greater than the revenue increment of the current re-insertion request at other insertion positions.

[0022] The target insertion position is randomly determined from the optimal insertion position and multiple suboptimal insertion positions based on a random perturbation factor;

[0023] Calculate the revenue increment when the current re-insertion request is inserted into the partial solution according to the target insertion position, obtain the revenue increment of the insertion operation, and calculate the revenue increment of the new transportation path created for the current re-insertion request.

[0024] Determine whether the revenue increment of the insertion operation is greater than the revenue increment of the new transportation route;

[0025] If the incremental revenue from the insertion operation is greater than the incremental revenue from the new transportation path, then the current request to be re-inserted is inserted into the partial solution according to the target insertion position to obtain a new partial solution.

[0026] The next request to be re-inserted is identified as the current request to be re-inserted. Based on the new partial solution, the step of calculating the incremental benefit of the current request to be re-inserted at different insertion positions in the partial solution and its subsequent steps are re-executed until all requests to be re-inserted are traversed and a new current solution is obtained.

[0027] Iterate through all request pairs located on different transportation paths in the new current solution;

[0028] Determine whether the request pair being traversed is recorded in the pre-built taboo list;

[0029] If the request pair being traversed is not recorded in the pre-built taboo list, then a cross-path swap operation is allowed to be performed on the request pair being traversed to generate a new solution, and the request pair being traversed is recorded in the pre-built taboo list, and an initial taboo length value is set for the request pair being traversed; the cross-path swap operation is the operation of swapping two request nodes in two different transport paths.

[0030] When performing a cross-path swap operation on the currently traversed request pair to generate a new solution, determine whether the new solution satisfies the preset time window constraint;

[0031] When the new solution does not meet the preset time window constraint, the total number of violations in the time window is quantified using a preset penalty function, and the total number of violations in the time window is converted into a penalty value based on a preset penalty coefficient. The penalty value is then added to the original net profit of the new solution to obtain the net profit of the new solution.

[0032] Determine whether the new solution is feasible and whether the net profit of the new solution is greater than the net profit of the current optimal solution; wherein, the current optimal solution is the solution with the highest net profit among all feasible solutions recorded in the current iteration.

[0033] If the new solution is the feasible solution and the net profit of the new solution is greater than the net profit of the current optimal solution, then the new solution is determined as the new current optimal solution, and the request pair traversal of the current iteration ends.

[0034] At the end of the current iteration, the tabu length values ​​of the existing request pairs in the tabu list are updated to obtain the updated tabu list, so that execution can begin in the next iteration based on the new current solution, the new current optimal solution, and the updated tabu list.

[0035] Optionally, the step of determining whether the currently traversed request pair is recorded in the pre-built taboo list further includes:

[0036] If the request pair being traversed is recorded in the pre-built taboo list, and the new solution generated when performing a cross-path exchange operation on the request pair being traversed is a feasible solution and the net profit of the new solution is greater than the net profit of the current optimal solution, then the taboo restrictions on the request pair being traversed are lifted, and a cross-path exchange operation is allowed to be performed on the request pair being traversed to generate a new solution. The new solution is then determined as the new current optimal solution, and the traversal of the request pair in the current iteration ends.

[0037] Optionally, the step of simulating the dynamic consumption of trailer resources during the scheduling execution process based on the ideal transportation scheduling scheme to identify shortage requests where trailer resources are scarce, obtaining a shortage request list, and generating a trailer relocation task set based on the shortage request list includes:

[0038] All container transport requests in the ideal transport scheduling scheme are sorted in ascending order according to the requested service start time.

[0039] According to the time order of the sorted container transport requests, the dynamic consumption of trailer resources during the scheduling and execution process of each container transport request is simulated in turn to dynamically update the inventory of trailers of different sizes in each trailer yard.

[0040] If the inventory quantity of trailers of the required size for the current simulated container transport request is negative, then the current simulated container transport request is marked as a shortage request due to a shortage of trailer resources, and a shortage request list is obtained.

[0041] For each shortage request in the shortage request list, generate a trailer repositioning task to supplement trailer resources, and obtain a corresponding trailer repositioning task set.

[0042] Optionally, for each shortage request in the shortage request list, a trailer repositioning task for supplementing trailer resources is generated, resulting in a corresponding set of trailer repositioning tasks, including:

[0043] For each shortage request in the shortage request list, query whether there is a candidate request in the ideal transportation scheduling scheme that meets preset constraints: wherein the preset constraints include the constraint that the completion time of the candidate request is earlier than the start time of the shortage request, the constraint that the container size required to be transported by the candidate request matches the container size required to be transported by the shortage request, and the constraint that the second identifier of the trailer operation information included in the candidate request indicates that the tractor will leave the trailer at the delivery point after the delivery is completed.

[0044] If there are multiple candidate requests that satisfy the preset constraints in the ideal transportation scheduling scheme, then the target candidate request with the earliest completion time is determined from the multiple candidate requests.

[0045] Calculate the time difference between the start time of the shortage request and the completion time of the target candidate request, and calculate the transportation time to transport the trailer from the delivery point of the target candidate request to the trailer yard required for the shortage request, and then from the trailer yard to the pickup point of the shortage request.

[0046] Determine whether the time difference is greater than the transportation time;

[0047] If the time difference is greater than the transportation time, a corresponding trailer repositioning task is generated.

[0048] The pickup point for the trailer repositioning task is the delivery point of the target candidate request, and the delivery point for the trailer repositioning task is either the trailer yard required for the shortage request or the pickup point of the shortage request. The time window for the trailer repositioning task is determined based on the completion time of the target candidate request, the start time of the shortage request, and the transportation time from the trailer yard to the pickup point of the shortage request.

[0049] Alternatively, if no candidate request that meets the preset constraints exists in the ideal transportation scheduling scheme, a trailer repositioning task is generated to allocate a trailer from an independent external resource that matches the size of the container required for the shortage request, and a preset penalty cost is added to the cost objective of the trailer repositioning scheduling model for the trailer repositioning task.

[0050] Optionally, corresponding model constraints are set during the process of constructing a transportation scheduling planning model based on the set of container transportation requests with the goal of maximizing net revenue;

[0051] The model constraints are determined based on the first trailer operation information and first container size information corresponding to the first container transport request, and the second trailer operation information and second container size information corresponding to the second container transport request. The determination is made on whether the tractor needs to go to the trailer yard to perform trailer collection operation, trailer return operation, or trailer exchange operation on its way from the delivery point of the first container transport request to the pickup point of the second container transport request. The model constraints are integrated into the transport scheduling planning model in the form of linear constraint expressions.

[0052] The first container transport request and the second container transport request are consecutive container transport requests.

[0053] Optionally, generating a container transportation and trailer resource scheduling scheme based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme includes:

[0054] The ideal transportation scheduling scheme and the trailer repositioning scheduling scheme are integrated on the timeline to generate a container transportation and trailer resource scheduling scheme that does not conflict with the use of tractor resources and trailer resources.

[0055] The container transportation and trailer resource scheduling scheme includes a tractor assignment table for recording the daily mileage of each tractor and a trailer scheduling tracking table for recording the movement trajectory of each trailer.

[0056] Secondly, the present invention also discloses a container transportation scheduling device, wherein the device comprises:

[0057] The request acquisition module is used to acquire a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether the tractor needs to carry a trailer of matching size to the pickup point, and a second identifier indicating whether the tractor should leave the trailer at the delivery point after the delivery is completed.

[0058] The first model building module is used to build a transportation scheduling planning model based on the set of container transportation requests with the goal of maximizing net revenue.

[0059] The first model solving module is used to solve the transportation scheduling planning model using a preset heuristic algorithm to obtain an ideal transportation scheduling scheme; the preset heuristic algorithm is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and taboo memory.

[0060] The relocation task generation module is used to simulate the dynamic consumption of trailer resources during the scheduling process based on the ideal transportation scheduling scheme, so as to identify the shortage requests where there is a shortage of trailer resources, obtain a shortage request list, and generate a trailer relocation task set based on the shortage request list.

[0061] The second model building module is used to build a trailer relocation scheduling model based on the trailer relocation task set with the goal of minimizing the total cost of trailer relocation;

[0062] The second model solving module is used to solve the trailer relocation and scheduling model to obtain the trailer relocation and scheduling scheme.

[0063] The scheduling scheme generation module is used to generate a container transportation and trailer resource scheduling scheme based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme.

[0064] Thirdly, the present invention discloses a terminal, comprising: a memory, a processor, and a container transport scheduling program stored in the memory and executable on the processor, wherein the container transport scheduling program, when executed by the processor, implements the steps of the container transport scheduling method as described above.

[0065] Fourthly, the present invention discloses a computer-readable storage medium storing a computer program that can be executed to implement the steps of the container transport scheduling method described above.

[0066] This invention provides a container transport scheduling method, apparatus, terminal, and medium. The container transport scheduling method includes: acquiring a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether a tractor needs to carry a trailer of matching size to a pickup point, and a second identifier indicating whether the tractor will leave the trailer at the pickup point after delivery; constructing a transport scheduling planning model based on the set of container transport requests with the objective of maximizing net revenue, and solving the transport scheduling planning model using a preset heuristic algorithm to obtain an ideal transport scheduling scheme; the preset heuristic algorithm is based on deconstruction and reconstruction of the Big Neighbor algorithm. An iterative optimization algorithm is used, which integrates domain search with local search carrying constraint relaxation and tabu memory. Based on the ideal transportation scheduling scheme, the dynamic consumption of trailer resources during the scheduling execution process is simulated to identify shortage requests where trailer resources are scarce, resulting in a shortage request list. A trailer relocation task set is generated based on the shortage request list. A trailer relocation scheduling model is constructed based on the trailer relocation task set, aiming to minimize the total cost of trailer relocation. The trailer relocation scheduling model is solved to obtain a trailer relocation scheduling scheme. A container transportation and trailer resource scheduling scheme is generated based on the ideal transportation scheduling scheme and the trailer relocation scheduling scheme. Therefore, this invention constructs and solves a transportation scheduling planning model that considers trailer operation information based on the acquired set of container transportation requests, obtaining an ideal transportation scheduling scheme with the goal of maximizing net revenue. Subsequently, the scheduling execution process is simulated, trailer yard resource shortages are dynamically identified, and a set of trailer repositioning tasks is generated. Then, a trailer repositioning scheduling model with the goal of minimizing the total cost of trailer repositioning is constructed and solved, obtaining a trailer repositioning scheduling scheme as a remedial scheduling. The ideal transportation scheduling scheme and the trailer repositioning scheduling scheme are merged to output a resource-feasible container transportation and trailer resource scheduling scheme. In other words, the technical solution of this invention, through a two-stage collaborative optimization framework and a preset hybrid heuristic algorithm, ensures the dynamic balance of trailer resources and the actual operability of the scheduling scheme, effectively solving the problem of infeasibility of planning caused by neglecting resource flow in traditional container scheduling transportation models, and realizing refined and intelligent scheduling management of the entire container land transportation chain. Attached Figure Description

[0067] Figure 1 This is a flowchart of a preferred embodiment of the container transportation scheduling method in this invention;

[0068] Figure 2 This is a schematic diagram of a specific trailer operation path that a tractor may perform between two consecutive requests, as disclosed in this invention.

[0069] Figure 3 This is a schematic diagram of a specific container transportation and trailer repositioning scheduling scheme disclosed in this invention, which includes four requests;

[0070] Figure 4 This is a functional principle block diagram of a preferred embodiment of the container transportation scheduling device in this invention;

[0071] Figure 5 This is a functional principle block diagram of a preferred embodiment of the terminal in this invention. Detailed Implementation

[0072] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0073] It is important to note that container shipping is a crucial link in the global logistics system, especially in multimodal transport. In land-based transport in port cities, tractor-trailers combine to complete the "last mile" transport from the terminal or yard to the customer's warehouse; this process is commonly referred to as container shipping. Although this link accounts for only 5% to 10% of the total transport distance, it can account for up to 30% of the total cost, and its operational efficiency directly impacts the profitability of the entire logistics chain.

[0074] Currently, optimization strategies for container shipping assume that tractor-trailers are always paired, and that trailers are not allowed to be left at the delivery point. However, in practice, customers often require shipping companies to provide empty trailers at the pickup point or leave fully loaded trailers after delivery for subsequent processing due to loading / unloading equipment or time constraints. This practical need for "trailer detachability" makes trailers a scarce resource that dynamically flows between the yard and the customer's location, easily leading to yard resource depletion and rendering pre-defined static scheduling plans unenforceable. In other words, while leaving trailers at the delivery point or dispatching new trailers from the trailer yard are important means of solving transportation scheduling problems in container logistics, leaving trailers at the delivery point may result in insufficient trailer availability at the trailer yard or increased transportation costs, which is detrimental to the feasibility and efficiency of container hauling operations.

[0075] Moreover, existing technologies mainly focus on tractor route optimization or empty container relocation. Even when the separation of tractor and trailer is involved, they fail to integrate and coordinate the constraints of "trailers being available and available for placement" with the subsequent necessary "dynamic relocation of trailer resources." This results in resource conflicts in the actual implementation of container transport scheduling schemes, or the need for high temporary relocation costs to remedy the situation, lacking robustness and economy in complex actual operating environments.

[0076] To this end, this application provides a container transportation scheduling scheme that can simultaneously optimize the dynamic balance of transportation routes and trailer resources, and coordinate the optimization of the actual needs of trailer separability and the dynamic repositioning of trailer resources, thereby improving port collection and distribution efficiency and reducing operating costs.

[0077] Please see Figure 1 , Figure 1 This is a flowchart of the container transportation scheduling method in this invention. For example... Figure 1 As shown, the container transportation scheduling method described in this embodiment of the invention includes:

[0078] Step S11: Obtain a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether the tractor needs to carry a trailer of matching size to the pickup point, and a second identifier indicating whether the tractor will leave the trailer at the delivery point after delivery.

[0079] In this embodiment, the container transport request set This includes all pending customer transportation requests within a defined planning period. It can be a single business day, from 0 to 24 hours. Furthermore, each request in the container shipping request set includes, in addition to trailer operation information, pickup point, delivery point, time window, container dimensions, transportation revenue, and transportation cost; that is, each request... It is a structured data object containing the following key attributes:

[0080] The pickup point indicates the specific location where the container needs to be picked up, which could be a dock, empty container yard, factory warehouse, etc. The delivery point indicates the destination where the container needs to be delivered, usually the customer's warehouse or a designated site. Indicates from arrive The transit time is a fixed parameter calculated based on geographical distance and average travel speed. Indicates a container shipping request Delivery point Container shipping request Pick-up point The empty driving time is crucial for planning the connection between different container transport requests for tractor units; The service window refers to the time range within which the customer allows loading and unloading operations to be performed, during which the tractor must... Then begin the service, and preferably at Completed previously; Indicates at the pickup point Loading service time (loading is loading containers onto trailers); For delivery point The unloading service time (unloading is the process of unloading the container from the trailer) may be shortened if the trailer is held in place. For container size markings, such as =0 indicates a 20-foot container. =1 indicates a 40-foot container, which determines the required trailer size; To complete the request Available transportation revenue; The direct transportation costs incurred in fulfilling request r; The first identifier is a binary variable. =1 indicates a pickup point No matching empty trailer is provided; the tractor must bring its own trailer of the same size. =0 indicates a pickup point An empty trailer has been provided; a tractor unit can proceed directly there. This is the second identifier, and also a binary variable. =1 indicates that the customer requires delivery at the specified point. When leaving a loaded trailer, the tractor-trailer must leave the trailer and container behind and then depart empty. =0 indicates that the customer does not leave the trailer, and the tractor must leave with the trailer (whether empty or fully loaded). These two indicators are derived from actual customer operation instructions and directly affect whether the tractor needs to visit the trailer yard to adjust trailer resources before and after performing the task.

[0081] It should be noted that other data involved in building a transportation scheduling planning model aimed at maximizing net revenue, such as the set of tractor resources, the set of tractor yards, the set of trailer yards, and related constraint parameters, can be automatically obtained through API integration with the enterprise transportation management system, warehouse management system, and customer order platform.

[0082] Among them, the collection of tractor resources Representing all available homogeneous tractor units, each tractor unit can only tow one trailer at a time; tractor unit yard collection. This indicates the base where tractor units are parked before the planning period begins and must be returned to after the planning period ends; each tractor unit yard... There is an initial number of available tractors. Trailer yard assembly This area is specifically designed for storing and allocating empty trailers of different sizes. Tractors collect, return, or exchange trailers here. Each yard... At the start of the planning period, initial inventory is set for each size of trailer. ; The planning time range is typically set to 0 to 24 hours; The fixed cost incurred for dispatching each tractor unit; For the tractor to travel through the arc ( , The unit distance cost or time cost; For when the tractor is performing from arrive During the trip, an extra detour was made to the trailer yard. The additional costs incurred in performing trailer operations are typically related to the distance traveled. (Extra time) is directly proportional.

[0083] Step S12: Construct a transportation scheduling planning model with the goal of maximizing net revenue based on the container transportation request set, and use a preset heuristic algorithm to solve the transportation scheduling planning model to obtain an ideal transportation scheduling scheme; the preset heuristic algorithm is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and taboo memory.

[0084] In this embodiment, the aim is to formulate a transportation task execution plan that maximizes the operator's total profit under the ideal condition of global balance of trailer resources. The implementation process is divided into two parts: mathematical modeling and algorithm solution. That is, a transportation scheduling planning model with the goal of maximizing net revenue is constructed based on the set of container transportation requests, and the ideal transportation scheduling scheme is obtained by solving the transportation scheduling planning model using a preset heuristic algorithm.

[0085] It should be noted that net revenue is the total transportation revenue minus the total cost, including fixed scheduling costs, operating costs, and additional costs incurred in performing trailer collection, return, or exchange operations. The transportation scheduling planning model is a first-stage mixed integer linear programming model constructed with the goal of maximizing net revenue, without considering the possibility of trailer shortages in trailer warehouses.

[0086] Specifically, the first-stage mixed-integer linear programming model is constructed into a directed graph. The relevant decision variables are as follows:

[0087] , indicating if arc ( , If the vehicle is driven by a tractor, the value is 1; otherwise, the value is 0.

[0088] This indicates if a request is made If the service is being provided, the value is 1; otherwise, it is 0.

[0089] Indicates the start of service request Time;

[0090] This indicates that if the tractor is at the node Drive to the node During the trip, we visited a trailer yard. If a trailer operation (collection, return, or exchange) is performed, the value is 1; otherwise, it is 0.

[0091] The objective function is to maximize net profit, as follows:

[0092] ;

[0093] in, This represents the total revenue from all service requests minus their direct transportation costs. This represents the total cost incurred by the tractor unit during empty runs between different task points. This refers to the detour costs incurred due to visiting trailer yards for trailer allocation. This represents the fixed scheduling cost.

[0094] Furthermore, in the process of constructing a transportation scheduling planning model based on a set of container transportation requests with the objective of maximizing net revenue, corresponding model constraints are set. These constraints involve determining, based on the first trailer operation information and first container size information corresponding to the first container transportation request, and the second trailer operation information and second container size information corresponding to the second container transportation request, whether the tractor unit needs to go to the trailer yard to perform trailer collection, trailer return, or trailer exchange operations during its journey from the delivery point of the first container transportation request to the pickup point of the second container transportation request. These constraints are integrated into the transportation scheduling planning model using linear constraint expressions. The first and second container transportation requests are consecutive container transportation requests. It is understandable that a series of constraints ensures the feasibility of the solution to the transportation scheduling planning model. These constraints are implemented through a set of constraints containing large... The linear inequalities accurately describe the logical judgments for all situations. The constraints of the transportation scheduling planning model include:

[0095] To constrain service requests, ensure that a request is served at most once, and if served, there must be an arc leaving its corresponding node;

[0096] For fleet size constraints, this means that the number of vehicles dispatched from each tractor yard does not exceed its initial inventory;

[0097] and To ensure flow balance, the number of vehicles flowing in equals the number of vehicles flowing out for both parking lot nodes and requesting nodes, thus forming a complete path.

[0098] constraint and The rule stipulates that each request requires a maximum of one operation on the trailer;

[0099] constraint Ensure that if no trailer is available at the pickup location, the tractor unit departing from the depot must go to the trailer yard to collect the trailer;

[0100] constraint The regulations stipulate that if the trailer is not left at the customer's delivery point, the tractor unit returning to the depot must return the trailer first.

[0101] constraint Four different trailer scenarios are described, representing the scenarios based on consecutive requests i and i. The trailer's operating and dimensional information forcibly determines the tractor's ability to complete tasks. Then go to Previously, was it necessary to visit the trailer yard (i.e.) See also Figure 2 As shown, it includes:

[0102] Scene 1: and The tractor unit is in Leave the trailer behind, and Trailers are provided; no operation is required. It should be 0.

[0103] Scene 2: and The tractor-trailer drove away. ,but Trailers are provided; the tractor must first go to the parking lot. Return the trailer. It should be 1.

[0104] Scene 3: and The tractor unit is in Leave the trailer behind, but Trailers are not provided; towing vehicles must first go to the parking lot. Collect trailers, It should be 1.

[0105] Scene 4: and The tractor-trailer drove away. ,and Trailers are not provided. If (If the dimensions are the same), the tractor can directly use the existing trailer to go there. It should be 0; if Then you must go to the parking lot. Exchange trailers, The value is 1.

[0106] Other constraints This ensures the feasibility of the route in terms of service start time, and also serves to eliminate sub-loop constraints;

[0107] constraint Ensure that all service requests are completed within the specified time frame;

[0108] A time window for the request was specified;

[0109] , and The domain of the decision variables is defined.

[0110] In this embodiment, a pre-defined heuristic algorithm is used to solve the MILP model (i.e., the transportation scheduling planning model) to generate the first-stage transportation scheduling scheme (i.e., the ideal transportation scheduling scheme). It should be noted that the MILP model is an NP-hard problem. For small-scale examples (e.g., 25 requests), solvers such as Gurobi can be used to directly find the optimal solution. However, for medium-to-large-scale practical problems (e.g., 200 to 400 requests), solvers cannot obtain a satisfactory solution within an acceptable timeframe. Therefore, an efficient hybrid large neighborhood search-tabu search algorithm is used, which is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and tabu memory.

[0111] Specifically, an initial feasible solution for the transportation scheduling planning model is constructed; starting from the initial feasible solution, a main iteration loop is executed until a preset iteration termination condition is met. The solution with the highest net benefit is selected from all feasible solutions recorded during the iteration process to obtain the ideal transportation scheduling scheme. The following sub-steps are executed in each main iteration:

[0112] First, a request is randomly selected from the current solution as a seed request, and the similarity between the seed request and other requests in the current solution is calculated. Based on the similarity, a first preset number of target requests are determined from the other requests, and the target requests are removed from the current solution to obtain a partial solution. The minimum similarity among the first preset number of target requests is not less than the highest similarity among the non-target requests in the other requests.

[0113] Then, the target request to be removed is identified as the request to be re-inserted, resulting in multiple requests to be re-inserted; each request to be re-inserted is traversed sequentially, and the currently traversed request to be re-inserted is identified as the current request to be re-inserted; the revenue increment of the current request to be re-inserted is calculated for different insertion positions in the partial solution; the optimal insertion position and a second preset number of suboptimal insertion positions are determined based on the revenue increment; the revenue increment of the current request to be re-inserted at each suboptimal insertion position is not greater than the revenue increment of the current request to be re-inserted at the optimal insertion position, and is greater than the revenue increment of the current request to be re-inserted at other insertion positions; the target insertion position is randomly determined from the optimal insertion position and multiple suboptimal insertion positions based on a random perturbation factor; the revenue increment when the current request to be re-inserted is inserted into the partial solution according to the target insertion position is calculated, the revenue increment of the insertion operation is obtained, and the revenue increment of the new transportation path created for the current request to be re-inserted is calculated; it is determined whether the revenue increment of the insertion operation is greater than the revenue increment of the new transportation path; if the revenue increment of the insertion operation is greater than the revenue increment of the new transportation path, the next step is to determine whether the revenue increment of the insertion operation is greater than the revenue increment of the new transportation path. When the revenue increment of the new transportation route is reached, the current request to be re-inserted is inserted into the partial solution according to the target insertion position to obtain a new partial solution. The next request to be re-inserted is identified as the current request to be re-inserted, and the steps to calculate the revenue increment of the current request to be re-inserted at different insertion positions in the partial solution and subsequent steps are re-executed based on the new partial solution until all requests to be re-inserted are traversed to obtain a new current solution. If the revenue increment of the insertion operation is not greater than the revenue increment of the new transportation route, it is further determined whether there are remaining vehicles in the current depot (e.g., whether the number of remaining vehicles is greater than zero). If there are remaining vehicles in the current depot, a new route operation is performed to add a new transportation route for the request to be re-inserted. If there are no remaining vehicles in the current depot, the current request to be re-inserted is still inserted into the partial solution according to the target insertion position to obtain a new partial solution. That is, it is determined whether there are enough vehicles available for adding a new route. If there are enough vehicles, a new route operation is performed; if there are not enough vehicles, an insertion operation is performed. The current vehicle yard refers to the tractor yard and trailer yard, and the remaining vehicle number refers to the remaining number of tractor vehicles and trailer vehicles.

[0114] Then, iterate through all request pairs located on different transport paths in the new current solution; determine whether the request pair being iterated through is recorded in the pre-built taboo list.

[0115] One specific scenario is as follows: when the request pair being traversed is not recorded in the pre-built taboo list, it is allowed to perform a cross-path swap operation on the request pair being traversed to generate a new solution, and the request pair being traversed is recorded in the pre-built taboo list, and an initial taboo length value is set for the request pair being traversed; the cross-path swap operation is the operation of swapping two request nodes in two different transport paths.

[0116] When performing a cross-path swap operation on the currently traversed request pair to generate a new solution, it is determined whether the new solution meets the preset time window constraint. If the new solution does not meet the preset time window constraint, the total number of violations in the time window is quantified using a preset penalty function, and the total number of violations in the time window is converted into a penalty value based on the preset penalty coefficient. The penalty value is then added to the original net profit of the new solution to obtain the net profit of the new solution. If the new solution meets the preset time window constraint, the original net profit of the new solution is directly determined as the net profit of the new solution.

[0117] Further, determine whether the new solution is feasible and whether its net profit is greater than the net profit of the current optimal solution. The current optimal solution is the solution with the highest net profit among all feasible solutions recorded in the current iteration. If the new solution is feasible and its net profit is greater than the net profit of the current optimal solution, then the new solution is determined as the new current optimal solution, and the traversal of request pairs in the current iteration ends. At the end of the current iteration, update the tabu length value of the existing request pairs in the tabu list to obtain the updated tabu list, so that execution can begin in the next iteration based on the new current solution, the new current optimal solution, and the updated tabu list.

[0118] Furthermore, if the new solution is not feasible or the net profit of the new solution is not greater than the net profit of the current optimal solution, then the next request pair is traversed, and the next request pair is determined as the current traversed request pair. The step of judging whether the current traversed request pair is recorded in the pre-constructed taboo list and its subsequent steps are re-executed until all request pairs located on different paths in the new current solution are traversed, or the new solution is feasible and the net profit of the new solution is greater than the net profit of the current optimal solution.

[0119] Another specific case is: if the request pair being traversed is recorded in the pre-built taboo list, and the new solution generated when performing a cross-path exchange operation on the request pair being traversed is a feasible solution and the net profit of the new solution is greater than the net profit of the current best solution, then the taboo restrictions on the request pair being traversed are lifted, and cross-path exchange operations are allowed to be performed on the request pair being traversed to generate a new solution. The new solution is directly determined as the new current best solution, and the traversal of the request pair in the current iteration ends.

[0120] For example, a predefined heuristic algorithm may include a destruction operator, a repair operator, and a 2-opt*-S local search operator. The execution of the predefined heuristic algorithm starts from an initial feasible solution constructed by the cheapest feasible insertion method. In each main iteration, destruction-repair and local search are performed sequentially.

[0121] Among them, the destruction operator can be the Shaw removal operator, which removes a batch of highly similar requests from the current solution based on request similarity; the repair operator can be the hybrid regret insertion operator, which aims to reinsert the removed requests into the current solution or create new transportation routes; the local search can be the 2-opt*-S operator, which refines the repaired solution. The algorithm continuously tracks the historical best solution and sets the number of iterations to stop as the termination condition.

[0122] Specifically, the core idea of ​​the Shaw removal operator is to remove similar requests from the current solution, thereby creating more optimization space for subsequent reorganization. Its similarity metric can comprehensively consider factors such as the geographical location of the request (distance to the delivery point), time window, and required trailer size. First, a seed request is randomly selected. Then, the similarity between the remaining requests in the solution and the seed request is calculated, and the requests with the highest similarity are removed sequentially. The request was partially resolved.

[0123] Furthermore, the hybrid regretful insertion operator re-inserts the removed request into the partial solution. It should be noted that traditional greedy insertion always chooses the insertion position with the largest increase in profit, which is prone to getting trapped in local optima, while the regretful strategy considers several suboptimal insertion positions in addition to the optimal insertion position, and randomly selects one of these suboptimal insertion positions for insertion.

[0124] This reduces the potentially huge cost of inserting the same request in the future. Furthermore, a hybrid strategy is employed, using a random factor to randomly select from the top few better positions with a certain probability, rather than choosing the optimal position, thus enhancing the diversity of the search. Simultaneously, the hybrid regret insertion operator also evaluates the possibility of creating a completely new route for each request.

[0125] The 2-opt*-S local search operator is a refined operator specifically designed for path problems with time windows. Unlike the traditional 2-opt (which swaps two edges in a path), it swaps the successor requests of two consecutive request pairs on two different paths. The 2-opt*-S local search operator swaps the requests in path A... Requests in path B Exchange, while keeping their respective predecessors and This approach effectively alters the path structure and makes it easier to maintain time window feasibility. Meanwhile, a taboo list is used during the search process to record recently swapped request pairs to avoid loops. Furthermore, a pre-defined contempt criterion allows for the acceptance of taboo moves that can produce better solutions. In other words, if a taboo request pair can produce a feasible solution, and the total profit of that feasible solution is better than the total profit of the current best solution, then the taboo restriction on that task pair is lifted, allowing the swap operation to be performed and the current best solution to be updated.

[0126] Furthermore, to help the algorithm escape local optima, the search process is allowed to temporarily accept solutions that slightly violate the time window constraint. This is achieved by introducing a penalty function. To quantify the path The sum of violations across all time windows is multiplied by a penalty coefficient and then added to the objective function. In the later stages of the algorithm or when outputting the final solution, it is necessary to ensure that =0 Therefore, the solution is completely feasible.

[0127] For example, if there are 10 tasks (i.e., container transport requests) that need to be transported, and a total of 3 tractors are available, the first preset number of removal operators can be 3, and the second preset number of insertion operators can be 2. The first preset number must be greater than or equal to the second preset number; otherwise, there will be insufficient tasks available for insertion. The preset taboo length can be 5, and the maximum number of iterations (i.e., the number of iteration stops) can be 2000. A task refers to a transport task, which involves transporting goods from the origin to the destination within a specified time window.

[0128] In solving the transportation scheduling planning model, an initial feasible solution is first generated. This initial feasible solution is then set as the current solution and the current optimal solution, and the total profit of the current solution and the current optimal solution is calculated respectively. The initial feasible solution is generated only once, but multiple iterations are performed. Based on the above basic parameters and the initial feasible solution, a cyclical iteration begins. In one iteration, the following sub-steps are executed:

[0129] The following removal operation is performed on the current solution: A task is randomly selected as the seed task. The similarity between other tasks and the seed task is calculated. After sorting the similarity scores in descending order, a first preset number (3) of tasks with the highest similarity are removed, yielding a partial solution. Simultaneously, the 3 removed tasks are designated as tasks to be re-inserted. The similarity score can be calculated based on the straight-line distance between the task destinations; the shorter the straight-line distance, the higher the similarity.

[0130] For partial solutions, the insertion operation is performed as follows: For the first task to be re-inserted, the revenue increment of the task to be re-inserted at different insertion positions is calculated; based on the revenue increment, the optimal insertion position and a second preset number (i.e., 2) of suboptimal insertion positions are determined; the revenue increment of the task to be re-inserted at each suboptimal insertion position is no greater than the revenue increment of the task to be re-inserted at the optimal insertion position, and is greater than the revenue increment of the task to be re-inserted at other insertion positions; one insertion position is randomly selected from the optimal insertion position and the two suboptimal insertion positions, and the insertion revenue increment is calculated; simultaneously, if there are still available vehicles, the revenue increment of the newly created transportation route for the task to be re-inserted is calculated; the insertion revenue increment and the revenue increment of the newly created transportation route are compared, and the one with the larger revenue increment is selected to perform the corresponding operation to complete the insertion operation of the first task to be re-inserted, obtaining a new partial solution. This process is repeated for all three tasks to be re-inserted, resulting in a new current solution.

[0131] It should be noted that if the new current solution is feasible, the total profit of the new current solution is calculated. If the total profit is greater than the total profit of the current optimal solution, the new current solution is set as the current optimal solution, and the total profit of the current optimal solution is updated.

[0132] Perform a cross-path swap operation (i.e., the 2-opt*-S local search operator) on the new current solution: The swap operation generates a new solution by exchanging two task nodes in two different paths. Specifically, iterate through all task pairs in the new current solution that are located on different paths according to their position in the paths. That is, first examine the first task of path 1 and the first task of path 2, i.e., task pair (1-1, 2-1), then examine the first task of path 1 and the second task of path 2, i.e., task pair (1-1, 2-2), and so on.

[0133] When the second task in path 1 and the fourth task in path 3 are swapped to obtain a new solution, if the new solution is feasible and its total profit is greater than the total profit of the current optimal solution, then the new solution is set as the new current optimal solution, the total profit of the current optimal solution is updated, and the traversal ends; otherwise, the next set of task pairs is traversed to perform cross-path swap operations.

[0134] While obtaining task pairs, a tabu list is constructed to record task pairs that have performed cross-path swap operations: (1-2, 3-4). The preset tabu length value for this task pair can be 5. If a task pair can produce a more feasible solution than the current optimal solution, then the task pair is not restricted by the tabu list. The tabu list is shown in Table 1.

[0135] Table 1

[0136]

[0137] At the end of the current iteration, the tabu values ​​of existing task pairs in the tabu list are updated, except for task pairs newly added in the current iteration, i.e. (1-2, 3-4). For example, if there is already a task pair (1-2, 2-3) in the tabu list with a tabu value of 3, its tabu value becomes 3-1=2 after the update. This means that after two more iterations, task 2 of path 1 and task 3 of path 2 can be used to perform cross-path swap operations.

[0138] The next iteration will be based on the latest current solution, the current optimal solution, and the taboo list, starting with the removal operation for the current solution. After 2000 iterations, the final current optimal solution is the ideal transportation scheduling scheme.

[0139] It is understood that the embodiments of this application obtain the optimal scheduling scheme for the first stage through iterative optimization of the aforementioned preset heuristic algorithm. (i.e., the ideal transportation scheduling scheme) is a set of tractor routes. Each route indicates the complete transportation path of the tractor starting from the yard, serving which requests in sequence, where to perform trailer operations, and finally returning to the yard.

[0140] Step S13: Based on the ideal transportation scheduling scheme, simulate the dynamic consumption of trailer resources during the scheduling execution process to identify shortage requests where trailer resources are scarce, obtain a shortage request list, and generate a trailer relocation task set based on the shortage request list.

[0141] It should be noted that the first phase plan It only guarantees the logical correctness of trailer operations on a single tractor path, but does not consider the dynamic balance of global trailer resources. Due to a large number of... The request is to leave the trailer at the customer's delivery point, trailer yard. The inventory will be continuously depleted, which may cause some subsequent requests to fail to execute due to the lack of matching trailers available.

[0142] In this embodiment, each repositioning task in the trailer repositioning task set includes a specified trailer transportation start point, destination, and a strict time window. This can be analyzed using a simulation program. This means simulating resource consumption and identifying request shortages.

[0143] Specifically, all container transport requests in the ideal transport scheduling scheme are sorted in ascending order according to the service start time of the requested requests; according to the time order of the sorted container transport requests, the dynamic consumption of trailer resources during the scheduling and execution process of each container transport request is simulated in turn to dynamically update the inventory quantity of trailers of different sizes in each trailer yard; when the inventory quantity of trailers of the required size for the current simulated container transport request is negative, the current simulated container transport request is marked as a shortage request with trailer resource shortage, and a shortage request list is obtained; for each shortage request in the shortage request list, a trailer relocation task to supplement trailer resources is generated, and a corresponding trailer relocation task set is obtained.

[0144] In this embodiment, for each shortage request in the shortage request list, a trailer repositioning task for supplementing trailer resources is generated, resulting in a corresponding set of trailer repositioning tasks. Specifically, this may include: for each shortage request in the shortage request list, querying whether there are candidate requests in the ideal transportation scheduling scheme that satisfy preset constraints: wherein, the preset constraints include the constraint that the completion time of the candidate request is earlier than the start time of the shortage request, the constraint that the container size required by the candidate request matches the container size required by the shortage request, and the constraint that the second identifier of the trailer operation information included in the candidate request indicates that the tractor will leave the trailer at the delivery point after delivery; when there are multiple candidate requests in the ideal transportation scheduling scheme that satisfy the preset constraints, then the task to be completed is determined from the multiple candidate requests. The earliest target candidate request is identified; the time difference between the start time of the shortage request and the completion time of the target candidate request is calculated, and the transportation time from the delivery point of the target candidate request to the trailer yard required for the shortage request, and then from the trailer yard to the pickup point of the shortage request, is calculated; it is determined whether the time difference is greater than the transportation time; if the time difference is greater than the transportation time, a corresponding trailer relocation task is generated; wherein, the pickup point of the trailer relocation task is the delivery point of the target candidate request, and the delivery point of the trailer relocation task is either the trailer yard required for the shortage request or the pickup point of the shortage request, and the time window of the trailer relocation task is determined based on the completion time of the target candidate request, the start time of the shortage request, and the transportation time from the trailer yard to the pickup point of the shortage request.

[0145] Furthermore, if there are no candidate requests that meet the preset constraints in the ideal transportation scheduling scheme, a trailer relocation task is generated to allocate trailers with the same container size as the required transport container from independent external resources, and a preset penalty cost is added to the cost objective of the trailer relocation scheduling model for the trailer relocation task.

[0146] For example, the input is (All routes and their timelines), initial inventory at each trailer yard Then All requests will begin service at their scheduled times. Sort in ascending order and scan each request in chronological order, i.e., check the trailer size required to serve the request. Check the origin of the trailer of the tractor unit that executed the request, based on the route and... The variable can determine which depot the tractor came from. The trailer was acquired from the parking lot. The size is The trailer inventory is decremented by 1. If the inventory becomes negative after decrementing by 1, the request is marked as a shortage request because at this point in the actual execution, there are no available trailers in the yard. Finally, a list of shortage requests is output. .against Each shortage request Generate a relocation task To supplement trailer resources, tasks can be generated in two ways, prioritizing internal allocation with lower costs. One method is to utilize trailers already stored at customer delivery points. In the middle, search for all in the request Plan start time The service has already been completed, and (The trailer was retained), and the container size... and Matching requests. These requests are placed with trailers at customer points, which are potential available resources; a completion time is selected from these. The earliest candidate request This ensures that the repositioning operation has the longest possible available time window; calculate the time difference: ; Calculate transportation time: That is, moving the trailer from the requested delivery point. Transported to the required parking lot , and then from Delivery to the shortage request pickup point The theoretical shortest time; if If there is sufficient time to complete the relocation, then a relocation task is generated. Its pickup point is Delivery point is (or directly) (depending on the modeling), the time window is Another approach is to urgently allocate trailers from external sources: if the previous method cannot find a candidate trailer that meets the time constraint, it is assumed that a trailer can be urgently allocated to the depot from an infinite pool of external resources. The generated task time window is... However, this requires paying a high penalty cost. This is to reflect the cost of such emergency dispatch. Ultimately, a set of relocation tasks is obtained. Each task has a clear start point, end point, time window, and cost attributes.

[0147] Step S14: Based on the trailer relocation task set, construct a trailer relocation scheduling model with the objective of minimizing the total cost of trailer relocation, and solve the trailer relocation scheduling model to obtain a trailer relocation scheduling scheme.

[0148] In this embodiment, the goal is to schedule the tractor unit to complete all generated repositioning tasks at minimal cost. Since the relocation task itself does not involve complex trailer operation logic, i.e., it does not need to determine whether the trailer is left behind or taken away, its model is relatively simple.

[0149] Specifically, the second-stage MILP model is constructed on another directed graph. G =(V ,A ) The decision variables in this stage are similar to those in the first stage, but simpler. In this stage, idle tractors that were available after the completion of the first stage can be used first. If these vehicles are used, their fixed scheduling costs will be... This cost may be cancelled or reduced, thus incentivizing resource reuse. The objective function is to minimize the total relocation cost, i.e.:

[0150] ;

[0151] The costs include transportation costs for the relocation task, empty run costs of the tractor, and fixed costs of calling up the tractor.

[0152] Furthermore, the feasibility of the solution to the trailer relocation scheduling model is ensured through a series of constraints, including:

[0153] constraint ,make sure All requests can be satisfied;

[0154] constraint and Ensure that the number of tractors used does not exceed the available number (including idle tractors and newly dispatched tractors). In addition, tractors not used in the first phase should be used first before new tractors are used in the second phase; otherwise, fines will be imposed.

[0155] constraint This ensures the conservation of flow;

[0156] constraint Ensure the request The service start time is no earlier than the previous request. The completion time eliminates the formation of sub-loops;

[0157] constraint Ensure all All service requests within the specified time frame are completed.

[0158] constraint It specifies the time window for requests;

[0159] constraint and It defines the domain of decision variables.

[0160] It should be noted that, because | |<<| The second-stage model (i.e., the trailer relocation and scheduling model) is much smaller in scale than the first-stage model (i.e., the transportation scheduling planning model). Therefore, commercial solvers such as Gurobi can be used to obtain the optimal or high-quality solution in a short time. After solving, the second-stage trailer relocation and scheduling scheme is obtained.

[0161] Step S15: Generate a container transportation and trailer resource scheduling scheme based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme.

[0162] In this embodiment, the ideal transportation scheduling scheme of the first stage and the trailer repositioning scheduling scheme of the second stage are combined to output a complete container transportation and trailer resource scheduling plan, that is, the independent schemes of the first two stages are integrated into a comprehensive operation plan that is globally feasible and resource-guaranteed.

[0163] Specifically, the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme are integrated on the timeline to generate a container transportation and trailer resource scheduling scheme that does not conflict with the use of tractor resources and trailer resources. The container transportation and trailer resource scheduling scheme includes a tractor assignment table for recording the daily journey of each tractor and a trailer scheduling tracking table for recording the movement trajectory of each trailer.

[0164] For example, see Figure 3 The example shown illustrates a container transport scheduling solution including trailer repositioning. This example fully demonstrates the collaborative mechanism between "tractor scheduling" and "trailer repositioning scheduling" within a two-stage scheduling framework, as well as core requirements such as trailer size matching, time windows, and closed-loop constraints. Specifically, the square node... Represents the tractor yard node, a triangular node. Represents a trailer yard node, a circular node. The diagram represents the pickup and delivery points for each request. Solid lines indicate the travel path of the tractor-trailer, dashed lines indicate the travel path of the tractor-trailer alone, and dotted lines indicate the travel path for trailer repositioning. This diagram mainly describes the execution flow of four sets of transport requests, starting from the tractor-trailer yard. Departure, passing through the trailer yard Once a matching trailer is found, the transport request is executed. and Further information will be provided based on the requested trailer details. Decide whether to detain trailers; when the trailer yard's inventory is insufficient, relocate trailers detained at the delivery location to the trailer yard to replenish inventory through trailer repositioning, ensuring the feasibility of subsequent requests, such as... The detained trailer was transferred to Service Request Simultaneously, all tractor units must eventually return to the tractor unit depot, forming a closed-loop scheduling mechanism. Integrating the ideal transportation scheduling scheme of the first phase (main transportation task) and the trailer repositioning scheduling scheme of the second phase (repositioning task) on the timeline requires ensuring that there are no conflicts in the use of tractor unit and trailer resources.

[0165] Specifically, regarding tractor resources: the second phase prioritizes using vehicles idled after the first phase, adding more only when necessary. In the integrated timetable, the same tractor cannot appear in two tasks simultaneously; this is ensured by the constraints of the second-phase model and the actual time schedule. Regarding trailer resources: the purpose of the relocation task is to supplement trailers for the main transportation task, and its completion time is strictly controlled before the corresponding shortage request begins. Therefore, in the complete container transportation and trailer resource scheduling plan, the supply of trailer resources is continuous and guaranteed in time.

[0166] Furthermore, a complete container transportation and trailer resource scheduling plan is a detailed set of time-based scheduling and transportation instructions, including at least a tractor assignment table and a trailer scheduling tracking table, and may also include key performance indicator reports and visualization charts.

[0167] The system includes: a tractor dispatch table recording the complete daily itinerary of each tractor (vehicle number), listed chronologically: departure from the depot, proceeding to the trailer depot (if necessary), executing a transport / relocation task sequence (including pick-up point, delivery point, and planned arrival / departure time for each task), and returning to the depot. A trailer dispatch tracking table shows the movement trajectory of each trailer (number), including: initial location, which tractor picked it up, when and where, which transport request it served, where it was left, which relocation task picked it up, and which depot it ultimately returned to. Key performance indicator reports include: total revenue, total cost (broken down), net revenue, total number of service requests, request completion rate, tractor utilization rate, and trailer turnover rate. Visualization charts include: a map display of tractor routes, and a graph showing resource inventory levels over time.

[0168] As can be seen, in this embodiment of the invention, a transportation scheduling planning model considering trailer operation information is constructed and solved based on the acquired set of container transportation requests to obtain an ideal transportation scheduling scheme with the goal of maximizing net revenue. Subsequently, the scheduling execution process is simulated to dynamically identify trailer yard resource shortages and generate a set of trailer repositioning tasks. Then, a trailer repositioning scheduling model with the goal of minimizing the total cost of trailer repositioning is constructed and solved to obtain a trailer repositioning scheduling scheme as a remedial scheduling. Finally, the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme are merged to output a resource-feasible container transportation and trailer resource scheduling scheme. That is, the technical solution of the present invention, through a two-stage collaborative optimization framework and a preset hybrid heuristic algorithm, ensures the dynamic balance of trailer resources and the actual operability of the scheduling scheme, effectively solving the problem of infeasibility of the plan caused by ignoring resource flow in the traditional container scheduling transportation model, and realizing refined and intelligent scheduling management of the entire chain of container land transportation.

[0169] The technical solution of this application can be integrated into the transportation management system of a logistics company as an intelligent dispatch engine. After operators import or the system automatically captures the next day's orders (request sets) daily, the engine is triggered to calculate, and the engine outputs an optimized dispatch plan within a few minutes to an hour; the plan can be automatically sent to the vehicle terminal to guide the driver's work; at the same time, the system can perform small-scale online rescheduling based on the actual work progress (such as delays).

[0170] In one embodiment, such as Figure 4 As shown, based on the above-described container transport scheduling method, the present invention also provides a container transport scheduling device, comprising:

[0171] The request acquisition module 11 is used to acquire a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether the tractor needs to carry a trailer of matching size to the pickup point, and a second identifier indicating whether the tractor should leave the trailer at the delivery point after the delivery is completed.

[0172] The first model construction module 12 is used to construct a transportation scheduling planning model based on the container transportation request set with the goal of maximizing net revenue.

[0173] The first model solving module 13 is used to solve the transportation scheduling planning model using a preset heuristic algorithm to obtain an ideal transportation scheduling scheme; the preset heuristic algorithm is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and taboo memory.

[0174] The relocation task generation module 14 is used to simulate the dynamic consumption of trailer resources during the scheduling process based on the ideal transportation scheduling scheme, so as to identify the shortage requests where there is a shortage of trailer resources, obtain a shortage request list, and generate a trailer relocation task set based on the shortage request list.

[0175] The second model construction module 15 is used to construct a trailer relocation scheduling model based on the trailer relocation task set with the goal of minimizing the total cost of trailer relocation.

[0176] The second model solving module 16 is used to solve the trailer relocation and scheduling model to obtain the trailer relocation and scheduling scheme.

[0177] The scheduling scheme generation module 17 is used to generate a container transportation and trailer resource scheduling scheme based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme.

[0178] Furthermore, it is worth noting that the working process of the container transportation scheduling device provided in this embodiment is the same as that of the container transportation scheduling method described above, so it will not be repeated here. For details, please refer to the working process of the container transportation scheduling method described above.

[0179] Figure 5 A schematic diagram of the structure of a terminal provided in an embodiment of this application. The terminal may include:

[0180] The memory 501, the processor 502, and the computer program stored on the memory 501 and capable of running on the processor 502.

[0181] When the processor 502 executes the program, it implements the container transportation scheduling method provided in the above embodiments.

[0182] Furthermore, the terminal also includes:

[0183] Communication interface 503 is used for communication between memory 501 and processor 502.

[0184] The memory 501 is used to store computer programs that can run on the processor 502.

[0185] Memory 501 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0186] If the memory 501, processor 502, and communication interface 503 are implemented independently, they can be interconnected via a bus to communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, only one line is used in the diagram, but this does not imply that there is only one bus or one type of bus.

[0187] Optionally, in a specific implementation, if the memory 501, processor 502, and communication interface 503 are integrated on a single chip, then the memory 501, processor 502, and communication interface 503 can communicate with each other through an internal interface.

[0188] Processor 502 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0189] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the container transport scheduling method described above.

[0190] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein.

[0191] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0192] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can read and execute instructions from and from an instruction execution system, apparatus or device).

[0193] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0194] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A container transport scheduling method, characterized in that, The method includes: Obtain a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether the tractor needs to carry a trailer of the same size to the pickup point, and a second identifier indicating whether the tractor should leave the trailer at the delivery point after the delivery is completed. Based on the set of container transport requests, a transport scheduling planning model is constructed with the goal of maximizing net revenue. The ideal transport scheduling scheme is obtained by solving the transport scheduling planning model using a preset heuristic algorithm. The preset heuristic algorithm is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and taboo memory. Based on the ideal transportation scheduling scheme, the dynamic consumption of trailer resources during the scheduling execution process is simulated to identify shortage requests where trailer resources are scarce, obtain a shortage request list, and generate a trailer relocation task set based on the shortage request list. Based on the set of trailer relocation tasks, a trailer relocation scheduling model is constructed with the goal of minimizing the total cost of trailer relocation, and the trailer relocation scheduling model is solved to obtain a trailer relocation scheduling scheme. A container transportation and trailer resource scheduling scheme is generated based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme. The process of simulating the dynamic consumption of trailer resources during the scheduling execution based on the ideal transportation scheduling scheme to identify trailer resource shortage requests, obtaining a shortage request list, and generating a trailer relocation task set based on the shortage request list includes: All container transport requests in the ideal transport scheduling scheme are sorted in ascending order according to the requested service start time. According to the time order of the sorted container transport requests, the dynamic consumption of trailer resources during the scheduling and execution process of each container transport request is simulated in turn to dynamically update the inventory of trailers of different sizes in each trailer yard. If the inventory quantity of trailers of the required size for the current simulated container transport request is negative, then the current simulated container transport request is marked as a shortage request due to a shortage of trailer resources, and a shortage request list is obtained. For each shortage request in the shortage request list, generate a trailer repositioning task to supplement trailer resources, and obtain a corresponding trailer repositioning task set; Furthermore, corresponding model constraints are set during the process of constructing a transportation scheduling planning model based on the set of container transportation requests with the goal of maximizing net revenue. The model constraints are determined based on the first trailer operation information and first container size information corresponding to the first container transport request, and the second trailer operation information and second container size information corresponding to the second container transport request. The determination is made on whether the tractor needs to go to the trailer yard to perform trailer collection operation, trailer return operation, or trailer exchange operation on its way from the delivery point of the first container transport request to the pickup point of the second container transport request. The model constraints are integrated into the transport scheduling planning model in the form of linear constraint expressions. The first container transport request and the second container transport request are consecutive container transport requests.

2. The container transport scheduling method according to claim 1, characterized in that, The process of solving the transportation scheduling planning model using a pre-defined heuristic algorithm to obtain an ideal transportation scheduling scheme includes: Construct an initial feasible solution for the transportation scheduling planning model; Starting from the initial feasible solution, execute the main iteration loop until the preset iteration termination condition is met. Select the solution with the highest net benefit from all feasible solutions recorded during the iteration process to obtain the ideal transportation scheduling scheme. The following sub-steps are executed in each main iteration: Randomly select a request from the current solution as a seed request, and calculate the similarity between the seed request and other requests in the current solution; Based on the similarity, a first preset number of target requests are determined from other requests, and the target requests are removed from the current solution to obtain a partial solution; the minimum similarity among the first preset number of target requests is not less than the highest similarity among the non-target requests in other requests. The target request to be removed is identified as a request to be re-inserted, resulting in multiple requests to be re-inserted; Each of the multiple re-insertion requests is sequentially traversed, and the currently traversed re-insertion request is determined as the current re-insertion request; Calculate the incremental revenue of the current re-insertion request at different insertion positions in the partial solution; The optimal insertion position and a second preset number of suboptimal insertion positions are determined based on the revenue increment; the revenue increment of the current re-insertion request at each of the suboptimal insertion positions is not greater than the revenue increment of the current re-insertion request at the optimal insertion position, but is greater than the revenue increment of the current re-insertion request at other insertion positions. The target insertion position is randomly determined from the optimal insertion position and multiple suboptimal insertion positions based on a random perturbation factor; Calculate the revenue increment when the current re-insertion request is inserted into the partial solution according to the target insertion position, obtain the revenue increment of the insertion operation, and calculate the revenue increment of the new transportation path created for the current re-insertion request. Determine whether the revenue increment of the insertion operation is greater than the revenue increment of the new transportation route; If the incremental revenue from the insertion operation is greater than the incremental revenue from the new transportation path, then the current request to be re-inserted is inserted into the partial solution according to the target insertion position to obtain a new partial solution. The next request to be re-inserted is identified as the current request to be re-inserted. Based on the new partial solution, the step of calculating the incremental benefit of the current request to be re-inserted at different insertion positions in the partial solution and its subsequent steps are re-executed until all requests to be re-inserted are traversed and a new current solution is obtained. Iterate through all request pairs located on different transportation paths in the new current solution; Determine whether the request pair being traversed is recorded in the pre-built taboo list; If the request pair being traversed is not recorded in the pre-built taboo list, then a cross-path swap operation is allowed to be performed on the request pair being traversed to generate a new solution, and the request pair being traversed is recorded in the pre-built taboo list, and an initial taboo length value is set for the request pair being traversed; the cross-path swap operation is the operation of swapping two request nodes in two different transport paths. When performing a cross-path swap operation on the currently traversed request pair to generate a new solution, determine whether the new solution satisfies the preset time window constraint; When the new solution does not meet the preset time window constraint, the total number of violations in the time window is quantified using a preset penalty function, and the total number of violations in the time window is converted into a penalty value based on a preset penalty coefficient. The penalty value is then added to the original net profit of the new solution to obtain the net profit of the new solution. Determine whether the new solution is feasible and whether the net profit of the new solution is greater than the net profit of the current optimal solution; wherein, the current optimal solution is the solution with the highest net profit among all feasible solutions recorded in the current iteration. If the new solution is the feasible solution and the net profit of the new solution is greater than the net profit of the current optimal solution, then the new solution is determined as the new current optimal solution, and the request pair traversal of the current iteration ends. At the end of the current iteration, the tabu length values ​​of the existing request pairs in the tabu list are updated to obtain the updated tabu list, so that execution can begin in the next iteration based on the new current solution, the new current optimal solution, and the updated tabu list.

3. The container transport scheduling method according to claim 2, characterized in that, The step of determining whether the currently traversed request pair is recorded in the pre-built taboo list also includes: If the request pair being traversed is recorded in the pre-built taboo list, and a new solution generated when performing a cross-path exchange operation on the request pair being traversed is a feasible solution and the net profit of the new solution is greater than the net profit of the current optimal solution, then the taboo restrictions on the request pair being traversed are lifted, and a cross-path exchange operation is allowed to be performed on the request pair being traversed to generate a new solution. The new solution is then determined as the new current optimal solution, and the traversal of the request pair in the current iteration ends.

4. The container transport scheduling method according to claim 1, characterized in that, For each shortage request in the shortage request list, a trailer repositioning task is generated to supplement trailer resources, resulting in a corresponding set of trailer repositioning tasks, including: For each shortage request in the shortage request list, query whether there is a candidate request in the ideal transportation scheduling scheme that meets preset constraints: wherein the preset constraints include the constraint that the completion time of the candidate request is earlier than the start time of the shortage request, the constraint that the container size required to be transported by the candidate request matches the container size required to be transported by the shortage request, and the constraint that the second identifier of the trailer operation information included in the candidate request indicates that the tractor will leave the trailer at the delivery point after the delivery is completed. If there are multiple candidate requests that satisfy the preset constraints in the ideal transportation scheduling scheme, then the target candidate request with the earliest completion time is determined from the multiple candidate requests. Calculate the time difference between the start time of the shortage request and the completion time of the target candidate request, and calculate the transportation time to transport the trailer from the delivery point of the target candidate request to the trailer yard required for the shortage request, and then from the trailer yard to the pickup point of the shortage request. Determine whether the time difference is greater than the transportation time; If the time difference is greater than the transportation time, a corresponding trailer repositioning task is generated. Wherein, the pickup point of the trailer repositioning task is the delivery point of the target candidate request, the delivery point of the trailer repositioning task is the trailer yard required for the shortage request or the pickup point of the shortage request, and the time window of the trailer repositioning task is determined based on the completion time of the target candidate request, the start time of the shortage request, and the transportation time from the trailer yard to the pickup point of the shortage request. Alternatively, if no candidate request that meets the preset constraints exists in the ideal transportation scheduling scheme, a trailer repositioning task is generated to allocate a trailer from an independent external resource that matches the size of the container required for the shortage request, and a preset penalty cost is added to the cost objective of the trailer repositioning scheduling model for the trailer repositioning task.

5. The container transport scheduling method according to any one of claims 1 to 4, characterized in that, The process of generating a container transportation and trailer resource scheduling scheme based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme includes: The ideal transportation scheduling scheme and the trailer repositioning scheduling scheme are integrated on the timeline to generate a container transportation and trailer resource scheduling scheme that does not conflict with the use of tractor resources and trailer resources. The container transportation and trailer resource scheduling scheme includes a tractor assignment table for recording the daily mileage of each tractor and a trailer scheduling tracking table for recording the movement trajectory of each trailer.

6. A container transport scheduling device, characterized in that, The device includes: The request acquisition module is used to acquire a set of container transport requests containing trailer operation information; the trailer operation information includes at least a first identifier indicating whether the tractor needs to carry a trailer of matching size to the pickup point, and a second identifier indicating whether the tractor should leave the trailer at the delivery point after the delivery is completed. The first model building module is used to build a transportation scheduling planning model based on the set of container transportation requests with the goal of maximizing net revenue. The first model solving module is used to solve the transportation scheduling planning model using a preset heuristic algorithm to obtain an ideal transportation scheduling scheme; the preset heuristic algorithm is an iterative optimization algorithm that combines large neighborhood search based on deconstruction and reconstruction with local search carrying constraint relaxation and taboo memory. The relocation task generation module is used to simulate the dynamic consumption of trailer resources during the scheduling process based on the ideal transportation scheduling scheme, so as to identify the shortage requests where there is a shortage of trailer resources, obtain a shortage request list, and generate a trailer relocation task set based on the shortage request list. The second model building module is used to build a trailer relocation scheduling model based on the trailer relocation task set with the goal of minimizing the total cost of trailer relocation; The second model solving module is used to solve the trailer relocation and scheduling model to obtain the trailer relocation and scheduling scheme. The scheduling scheme generation module is used to generate a container transportation and trailer resource scheduling scheme based on the ideal transportation scheduling scheme and the trailer repositioning scheduling scheme. The relocation task generation module is specifically used for: All container transport requests in the ideal transport scheduling scheme are sorted in ascending order according to the requested service start time. According to the time order of the sorted container transport requests, the dynamic consumption of trailer resources during the scheduling and execution process of each container transport request is simulated in turn to dynamically update the inventory of trailers of different sizes in each trailer yard. If the inventory quantity of trailers of the required size for the current simulated container transport request is negative, then the current simulated container transport request is marked as a shortage request due to a shortage of trailer resources, and a shortage request list is obtained. For each shortage request in the shortage request list, generate a trailer repositioning task to supplement trailer resources, and obtain a corresponding trailer repositioning task set; Furthermore, corresponding model constraints are set during the process of constructing a transportation scheduling planning model based on the set of container transportation requests with the goal of maximizing net revenue. The model constraints are determined based on the first trailer operation information and first container size information corresponding to the first container transport request, and the second trailer operation information and second container size information corresponding to the second container transport request. The determination is made on whether the tractor needs to go to the trailer yard to perform trailer collection operation, trailer return operation, or trailer exchange operation on its way from the delivery point of the first container transport request to the pickup point of the second container transport request. The model constraints are integrated into the transport scheduling planning model in the form of linear constraint expressions. The first container transport request and the second container transport request are consecutive container transport requests.

7. A terminal, characterized in that, include: The container transport scheduling program is stored in the memory and can run on the processor, wherein when executed by the processor, the container transport scheduling program implements the steps of the container transport scheduling method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that can be executed to implement the steps of the container transport scheduling method as described in any one of claims 1 to 5.

Images