An emergency material transportation optimization scheduling method based on a multimodal transport network

Abstract

The application discloses an emergency material transportation optimization scheduling method based on a multimodal transport network, and belongs to the field of logistics management. The method comprises the following steps: step 1, a multimodal transport network for emergency material transportation is built; step 2, a set of emergency material transportation tasks in a planning period is input; step 3, a multi-objective shortest path algorithm is used to generate multimodal transport paths between starting and ending points of each task in the multimodal transport network and the set of emergency material transportation tasks, a random disturbance method is used to change weight coefficients of optimization objectives, and a set of alternative paths is constructed; and step 4, a target function with the shortest completion time of the overall transportation plan as an optimization objective and constraint conditions of the target function are established according to the set of alternative paths, an adaptive large-scale neighborhood search algorithm is used to solve the target function, and an optimal overall scheduling scheme is obtained. The application makes the emergency material transportation scheme more reasonable and efficient, and provides reliable technical support for improving the management level of emergency material transportation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of logistics management, and in particular to an emergency material transportation optimization scheduling method based on a multimodal transport network. BACKGROUND

[0002] Emergency material transportation refers to a special logistics activity for emergency support after serious natural disasters, public health emergencies and other emergencies. Compared with traditional logistics transportation, the particularity of emergency material transportation requires the system to respond quickly and schedule efficiently in emergency situations. In recent years, the multimodal transport network has developed rapidly, with wide coverage, strong time efficiency and high resource integration, providing more choices and opportunities for emergency material transportation, and can cope with more complex and variable transportation demand and resource allocation problems.

[0003] At present, although there are some methods for emergency material transportation scheduling at home and abroad, most of them are still limited to single transportation mode, and lack sufficient consideration of multimodal transport. In addition, existing methods often ignore unexpected situations that may occur during transportation, such as: stations, routes cannot be used due to sudden disasters, etc. Therefore, in order to better cope with the demand for material transportation in emergency situations, an emergency material transportation optimization scheduling method based on a multimodal transport network is needed, which should consider the characteristics of different transportation modes, and take the availability of stations, roads, etc. into account as constraints, to realize the rapid deployment and optimal allocation of transportation resources. At the same time, this method should also fully consider factors such as the transportation distance of materials, traffic congestion, etc., to ensure that the transportation task of materials is completed within a limited time.

[0004] Multimodal transport: the transportation process completed by two or more vehicles in conjunction with each other is called multimodal transport.

[0005] Shortest path: the path with the minimum sum of weights on each edge from a node to another node along the edges of the graph is called the shortest path.

[0006] Adaptive Large Neighborhood Search (ALNS): a meta-heuristic algorithm that assigns weights to the destruction and repair operators based on large neighborhood search, so that the algorithm can automatically select the better operators to improve the solution, so as to obtain a better solution within a reasonable running time. The algorithm also uses the Metropolis acceptance criterion to avoid falling into local optimum. SUMMARY

[0007] The present application aims at the deficiencies in the prior art, and provides an emergency material transportation optimization scheduling method based on a multimodal transport network, which optimizes the transportation path and scheduling plan of the emergency material transportation task.

[0008] To achieve the above object, the present application adopts the following technical solutions:

[0009] An emergency material transportation optimization scheduling method based on a multimodal transport network, comprising:

[0010] Step 1, building a multimodal transport network for emergency material transportation, wherein the multimodal transport network comprises a hierarchical topology structure, station loading and unloading capacity and road section passing capacity;

[0011] Step 2, inputting a set of emergency material transportation tasks in a planning period, wherein the task set comprises a task starting point, a task ending point, a task transportation quantity, an earliest starting time, a latest delivery time and a task priority;

[0012] Step 3, generating multimodal transport paths between each task starting and ending point by using a multi-objective shortest path algorithm for the multimodal transport network and the tasks in the set of emergency material transportation tasks, changing the weight coefficient of the optimization objective by using a random disturbance method, and constructing a set of alternative paths;

[0013] Step 4, establishing an objective function with the shortest completion time of the overall transportation plan as the optimization objective and the constraint condition of the objective function according to the set of alternative paths, and solving the objective function by using an adaptive large-scale neighborhood search algorithm to obtain an optimal overall scheduling scheme.

[0014] To optimize the above technical solutions, the following specific measures are adopted:

[0015] Further, in step 1, the multimodal transport network for emergency material transportation is specifically:

[0016] The multimodal transport network is represented by a directed graph G=(V,E), wherein the station set is V={v0,v1,…,vn}, and vn represents the nth station; the road section set is E={(vi,vi+1):vi,vi+1∈V,i≠j}, and vi,vi+1 represents the station vi and vi+1. n n vi+1 represents the station vi and vi+1. i j vi+1 represents the station vi and vi+1. i j vi+1 represents the station vi and vi+1. i ​​​With site v j Connection relationships between them; subgraph G f =(V f E f () represents the hierarchical transportation network for different modes of transport, f∈F:={road, rail, waterway, air}, V f and E f Let V represent the set of stations and the set of road segments in mode f, respectively, and let F represent the set of modes of transport. Let V = V1∪V2∪…∪V |F| E = E1∪E2∪…∪E |F| ∪E t E t V represents the set consisting of all transfer routes. |F| E represents the set of stations for the last mode of transportation. |F| This represents the set of road segments that represent the last mode of transportation.

[0017] For site v i ∈V, This indicates the site v i The maximum loading capacity per unit time. This indicates the site v i The maximum unloading capacity per unit time; for general road sections (v i ,v j )∈E\E t , l i,j c represents the travel distance of a road segment. i,j Indicates the travel time of a road segment, h i,j This indicates the maximum traffic capacity of a road segment per unit of time; for transfer segments (v... i ,v j )∈E t , This indicates the transfer rate.

[0018] Furthermore, step 3 specifically includes:

[0019] Step 3.1: Select four indicators—route length, route time, number of transshipments, and transportation cost—as the optimization objectives for multimodal transport routes. Use the time value coefficient to normalize the optimization objectives across different dimensions, addressing the origin and destination points for each transport pair. The weight coefficients of the optimization objective are w1, w2, w3, and w4, respectively, satisfying ∑ i w i =1, i = 1, 2, 3, 4 and w i ≥0, randomly generate a set of weight coefficient combinations for the optimization objective.

[0020] The Dijkstra algorithm on the directed graph is used to calculate the optimal path under the current condition as the multimodal transport path;

[0021] Step 3.2, for the generated multimodal transport path, it is judged whether the following criteria are met: whether the path length is less than the allowed maximum value, whether the path time is less than the allowed maximum value, and whether the number of transfer times is less than the allowed maximum value, if all criteria are met, the generated multimodal transport path is added to the candidate path set;

[0022] Step 3.3, a random disturbance is applied to the weight coefficient combination, and the Dijkstra algorithm on the directed graph is used to calculate the optimal path under the current condition to generate a new multimodal transport path;

[0023] Step 3.4, steps 3.3.2 and 3.3.3 are repeated, and the repeated paths in the candidate path set are deleted until the number of paths in the candidate path set reaches the requirement.

[0024] Further, step 4 includes:

[0025] Step 4.1, a target function is established with the shortest completion time of the overall transport plan as the optimization objective, and the formula is as follows:

[0026]

[0027] Wherein, z represents the completion time of the overall transport plan, t r represents the start execution time of task r, 0-1 decision variable θ p represents whether p in the candidate path set is selected as the transport path of task r, if selected, it takes 1, otherwise 0, P(r) represents the candidate path set of task r, represents the transport time of path p; R represents the overall candidate path set;

[0028] The constraint condition of the target function is as follows:

[0029] Constraint condition 1: ensure that each task is assigned at least one path, and the number of paths assigned to each task is at most P num , which is expressed by formula as:

[0030]

[0031] In the formula, θ p represents a 0-1 decision variable;

[0032] Constraint condition 2:

[0033]

[0034] In the formula, δ is an integer variablep q represents the transport volume carried by path p. r This represents the transportation requirements of task r;

[0035] Constraint 3: Within each unit time window, the loading volume at any station shall not exceed the upper limit of the loading capacity, as expressed by the formula:

[0036]

[0037] In the formula, Indicates whether path p is at site v i To perform the loading activity, the 0-1 variable μ p,k Indicates whether path p is loaded within time window k; V is the set of sites; This indicates the maximum loading capacity of the site per unit of time; K represents the overall planning time window.

[0038] Constraint 4: Within each unit time window, the unloading volume of any site shall not exceed the upper limit of its unloading capacity, as expressed by the formula:

[0039]

[0040] In the formula, Indicates whether path p is at site v i To perform the uninstallation activity, the 0-1 variable v p,k Indicates whether path p is loaded within time window k; This indicates the site v i The maximum unloading capacity per unit time;

[0041] Constraint 5: Within each unit time window, the traffic volume of any road segment shall not exceed the upper limit of its capacity, as expressed by the formula:

[0042]

[0043] In the formula, parameter b i,j,p This indicates that path p is a route segment (v) i ,v j ), 0-1 variable ξ p,k Indicates whether path p is transported within time window k; E is the set of road segments;

[0044] Constraint 6:

[0045] θ p ,μ p,k ,v p,k ,ξ p,k ∈{0,1}

[0046] Constraint 7:

[0047] tr ,δ p ≥0

[0048] Step 4.2, solving the objective function by using an adaptive large-scale neighborhood search algorithm to obtain an optimal overall scheduling scheme.

[0049] Further, the solving the objective function by using an adaptive large-scale neighborhood search algorithm specifically is:

[0050] Step 4.2.1, for each transportation task, randomly selecting several initial paths from the set of alternative paths and randomly setting the transportation volume borne by the path obtaining the starting transportation time of each task calling a feasible solution construction method to obtain an initial solution s0=(x0,y0,t0); x0 represents the initial solution of the transportation path, y0 represents the initial solution of the transportation volume, and t0 represents the initial solution of the transportation time;

[0051] Step 4.2.2, initializing the optimal solution s * and the current solution s as the initial solution s0, setting the initial temperature, the final temperature, the cooling rate and the maximum number of iterations in the Metropolis acceptance criterion, and initializing the weights p of the destruction operator and the repair operator to the same value;

[0052] Step 4.2.3, according to the selection weights p of the destruction operator and the repair operator, selecting a pair of destruction operator and repair operator from the neighborhood set according to the roulette principle to perform neighborhood search operation, and calling a feasible solution construction method to generate a new solution representing the current solution of the transportation path, representing the current solution of the transportation volume, representing the current solution of the transportation time;

[0053] Step 4.2.4, comparing the new solution with the current solution s, if the new solution is better than the current solution, replacing the current solution s with the new solution , otherwise, judging whether to accept the new solution according to the Metropolis acceptance criterion, and updating the historical optimal solution s * ;

[0054] Step 4.2.5, scoring the destruction operator and the repair operator by collecting the performance of different destruction operators and repair operators in the past 100 iterations, updating the selection weights p of the destruction operator and the repair operator, and the scoring mechanism includes: (a) a new global optimal solution is obtained in the last iteration, (b) a solution is obtained in the last iteration that has not been found before, but the objective function value of the solution is worse than the current solution, (c) a solution is obtained in the last iteration that has not been found before, but the objective function value of the solution is worse than the current solution, but the solution is accepted;

[0055] Step 4.2.6, steps 4.2.3 to 4.2.5 are repeatedly executed until the iteration number reaches the maximum iteration number or the temperature reaches the minimum temperature, and the global optimal solution is output as s * .

[0056] Further, the feasible solution construction method refers to determining the solution of the transportation path and the solution of the transportation quantity, and then calculating the starting transportation time of each task according to the greedy strategy under the premise of meeting various transportation constraints, so as to construct a complete feasible solution, and the specific process is as follows:

[0057] Step 4.2.1.1, sort all tasks in descending order of task priority, while meeting the constraint condition of the sequence relationship between tasks, and obtain a queue List;

[0058] Step 4.2.1.2, initialize network resource occupation information, the loading and unloading capacity of the station, the passing capacity of the road section and the remaining capacity of other related transportation resources in each time window are all set to the default maximum value;

[0059] Step 4.2.1.3, obtain the first task of List, calculate the available resources of the task according to the path involved in the task and the current network resource occupation, and set the starting transportation time of each path in the task according to the greedy strategy, that is, execute the task as early as possible, and remove the current task from List after completion; repeat step 4.2.1.3 until the queue List is empty.

[0060] Further, the destruction operator refers to deleting a part of the solution of the task from the feasible solution according to certain rules, and the repair operator refers to resetting the solution of the deleted task according to certain rules, and a pair of destruction-repair operators defines a neighborhood structure, and the destruction operator and the repair operator are as follows:

[0061] Destruction operator 1: randomly select 10% of the tasks and delete their solutions;

[0062] Destruction operator 2: randomly select 20% of the tasks and delete their solutions;

[0063] Destruction operator 3: select the task with the latest completion time and delete his solution;

[0064] Repair operator 1: for all deleted tasks, randomly reset a number of initial paths from the alternative path set, and randomly reset the transportation quantity borne by the path;

[0065] Repair operator 2: for all deleted tasks, keep the original transportation path unchanged, and randomly reset the transportation quantity borne by the path.

[0066] The beneficial effects of the present application are:

[0067] The present application is based on the emergency material transportation problem after the occurrence of an emergency, and designs an emergency material transportation optimization scheduling method based on a multimodal transport network, solves the problems of low transportation efficiency of a single transportation mode and lack of collaborative organization between multiple transportation modes; a high-efficiency integrated multimodal transport network for emergency material transportation is built by comprehensively considering the hierarchical topological structure of multiple transportation modes and the transfer relationship between them; a multi-objective shortest path algorithm and a random disturbance method are used to generate multimodal transport paths, and a high-quality candidate path set is constructed; an adaptive large-scale neighborhood search algorithm is creatively used to solve the emergency material transportation optimization scheduling problem, so that the emergency material transportation scheme is more reasonable and efficient, and reliable technical support is provided for improving the management level of emergency material transportation. BRIEF DESCRIPTION OF DRAWINGS

[0068] Figure 1 The overall flowchart of the method proposed in the present application. DETAILED DESCRIPTION

[0069] The present application will now be described in further detail with reference to the accompanying drawings.

[0070] As shown in Figure 1 , an emergency material transportation optimization scheduling method based on a multimodal transport network includes the following steps:

[0071] Step 1, build a multimodal transport network for emergency material transportation, which contains hierarchical topological structure, station loading and unloading capacity, road section passing capacity, etc.

[0072] As the basic geographic data for emergency material transportation optimization scheduling, the multimodal transport network should have the following characteristics: 1) contains a hierarchical transport topological network of all transportation modes such as highways, waterways, railways and aviation, and has certain connectivity between various transportation modes, and can be transferred at some nodes; 2) the network topological structure is relatively simple, and the overall data volume is controlled within a calculable range; 3) considering the complex and variable characteristics of emergency material transportation, the network needs to be adjustable to respond to changes in the transportation environment.

[0073] The multimodal transport network is represented by a directed graph G=(V,E), where the station set is V={v0,v1,…,v n}, the road section set is E={(v i ,v j ):v i ,v j ∈V,i≠j}, representing the connection relationship between stations. The subgraph G f =(V f ,E f() represents the hierarchical transportation network for different modes of transport, f∈F:={road, rail, waterway, air}, V f and E f Let V represent the set of stations and the set of road segments in mode f, respectively, and let F represent the set of modes of transport. Let V = V1∪V2∪…∪V |F| E = E1∪E2∪…∪E |F| ∪E t E t V represents the set consisting of all transit (virtual) routes. |F| E represents the set of stations for the last mode of transportation. |F| This represents the set of road segments that use the last mode of transportation.

[0074] Assuming the minimum transport unit for emergency supplies is a 20-foot container (TEU), it is assumed that different types and quantities of supplies have been pre-packed and boxed. For site v i ∈V, This indicates the maximum loading capacity of a site per unit of time. This indicates the maximum unloading capacity of a station per unit of time; for general road sections (v i ,v j )∈E\E t , l i,j c represents the travel distance of a road segment. i,j Indicates the travel time of a road segment, h i,j This indicates the maximum traffic capacity of a road segment per unit of time; for transfer segments (v... i ,v j )∈E t , This indicates the transfer rate, which is the time required for a single transport unit to complete a transfer.

[0075] Step 2: Input the set of emergency supplies transportation tasks for the planned time period;

[0076] Task data is primarily determined by the relevant departments responsible for the overall management of emergency supplies transportation. Any transportation task r includes the task origin. Mission End Task transportation volume q r Earliest start time Latest delivery time Information such as task priority. Furthermore, different transportation tasks may have a sequential relationship; for example, a task may require the completion of its preceding task before it can begin execution.

[0077] Step 3, for the multi-modal transport network and the tasks in the set of emergency material transport tasks, a multi-objective shortest path algorithm is used to generate multi-modal transport paths between the start and end points of each task, the weight coefficients of the optimization objectives are changed by a random disturbance method, and a set of alternative paths is constructed;

[0078] The specific steps include:

[0079] Step 3.1, select path length, path time, number of transfers, and transportation cost as the optimization objectives of the multi-modal transport path, normalize the optimization objectives of different dimensions by using the time value coefficient, and for each pair of start and end points appearing in each transport task The weight coefficients of the optimization objectives are w1, w2, w3, and w4, which satisfy ∑ i w i =1, i=1,2,3,4 and w i ≥0, a set of weight coefficient combinations of the optimization objectives is randomly generated

[0080] The Dijkstra algorithm on the directed graph is used to calculate the optimal path under the current conditions as the multi-modal transport path;

[0081] Step 3.2, for the generated multi-modal transport path, judge whether the following standards are met: whether the path length is less than the allowed maximum value, whether the path time is less than the allowed maximum value, whether the number of transfers is less than the allowed maximum value, and other artificial constraints (such as passing through stations, avoiding stations, etc.). If all the standards are met, the generated multi-modal transport path is added to the set of alternative paths; the purpose of constructing the set of alternative paths is to provide different paths between the same start and end points for the emergency material transport unified scheduling method involved in step (4), thereby expanding the path selection range and avoiding the overuse of some key transport stations and transport lines to optimize the overall transport efficiency.

[0082] Step 3.3, apply random disturbance to the weight coefficient combination, and use the Dijkstra algorithm on the directed graph to calculate the optimal path under the current conditions to generate a new multi-modal transport path;

[0083] Step 3.4, repeat steps 3.3.2 and 3.3.3, and delete the duplicate paths in the set of alternative paths until the number of paths in the set of alternative paths reaches the requirement.

[0084] Step 4, according to the set of alternative paths constructed in step 3, establish a target function with the shortest completion time of the overall transport plan as the optimization objective, consider the constraints of transport resources and task sequence, use the adaptive large-scale neighborhood search algorithm to solve the emergency material transport optimization scheduling problem, and obtain the optimal overall scheduling scheme;

[0085] The specific steps include:

[0086] Step 4.1, constructing an optimized scheduling model: the emergency material transportation optimization scheduling needs to determine the transportation path for each transportation task, and reasonably arrange the time sequence plan on the fixed transportation path, so as to complete the overall transportation plan in the shortest time.

[0087] The objective function is established with the shortest completion time of the overall transportation plan as the optimization objective, and the formula is as follows:

[0088]

[0089] Wherein, the real number decision variable t r represents the starting execution time of task r, the 0-1 decision variable θ p represents whether p in the alternative path set is selected as the transportation path of task r, 1 is selected and 0 is not selected, represents the transportation time of path p, and R represents the overall alternative path set. The completion time of the last task is the completion time of the overall transportation plan.

[0090] Constraint conditions:

[0091]

[0092]

[0093]

[0094]

[0095]

[0096] θ p ,μ p,k ,v p,k ,ξ p,k ∈{0,1}(7)

[0097] t r ,δ p ≥0 (8)

[0098] Wherein, constraint (2) ensures that each task is assigned at least one path, and in order to ensure the implementability of the overall transportation plan, the number of paths assigned to each task is at most P num ; θ p represents a 0-1 decision variable; constraint (3) ensures that the transportation demand q r of each task is guaranteed, and the integer variable δ pdenotes the transportation volume of path p; constraint (4) ensures that the loading volume of any station in each unit time window does not exceed the upper limit of loading capacity, parameter denotes whether path p passes through station v i carries out loading activity, 0-1 variable μ p,k denotes whether path p carries out loading activity in time window k; V is a set of stations; denotes the upper limit of loading capacity of a station in a unit time; K denotes the total planning time window; constraint (5) ensures that the unloading volume of any station in each unit time window does not exceed the upper limit of unloading capacity, parameter denotes whether path p passes through station v i carries out unloading activity, 0-1 variable v p,k denotes whether path p carries out loading activity in time window k; denotes the upper limit of unloading capacity of station v i in a unit time; constraint (6) ensures that the traffic volume of any road section in each unit time window does not exceed the upper limit of traffic capacity, parameter b i,j,p denotes whether path p is a passing road section (v i ,v j ), 0-1 variable ξ p,k denotes whether path p transports in time window k; E is a set of road sections; constraints (7) and (8) define the value range of the decision variable.

[0099] Step 4.2, solving the optimization scheduling model: further, in this multi-modal transport network-based emergency material transportation optimization scheduling method, the adaptive large-scale neighborhood search algorithm is used to solve the optimization scheduling model described in step 4.1, and the specific process is as follows:

[0100] Step 4.2.1, for each transportation task, randomly select several initial paths from the set of alternative paths and randomly set the transportation volume of the path obtain the start transportation time of each task call the feasible solution construction method to obtain the initial solution s0=(x0,y0,t0); x0 represents the initial solution of the transportation path, y0 represents the initial solution of the transportation volume, and t0 represents the initial solution of the transportation time;

[0101] Step 4.2.2, initialize the optimal solution s * and the current solution s are the initial solution s0, the initial temperature, the final temperature, the cooling rate and the maximum iteration number in the Metropolis acceptance criterion are set, and the weights of the destruction operator and the repair operator are initialized to the same value;

[0102] Step 4.2.3, according to the selection weight p of the destruction operator and the repair operator, a pair of destruction operator and repair operator are selected from the neighborhood set according to the roulette wheel principle to perform neighborhood search operation, and the feasible solution construction method is called to generate a new solution denotes the current solution of the transportation path, denotes the current solution of the transportation volume, denotes the current solution of the transportation time;

[0103] Step 4.2.4, compare the new solution with the current solution s, if the new solution is better than the current solution, replace the current solution s with the new solution , otherwise, judge whether to accept the new solution according to the Metropolis acceptance criterion, and update the historical optimal solution s * ;

[0104] Step 4.2.5, score the destruction operator and the repair operator by collecting the performance of different destruction operators and repair operators in the past 100 iterations, update the selection weight p of the destruction operator and the repair operator, and the scoring mechanism includes: (a) a new global optimal solution is obtained in the last iteration, (b) a solution is obtained in the last iteration, but the objective function value of the solution is worse than the current solution, (c) a solution is obtained in the last iteration, but the objective function value of the solution is worse than the current solution, but the solution is accepted;

[0105] Step 4.2.6, repeat steps 4.2.3 to 4.2.5 until the number of iterations reaches the maximum number of iterations or the temperature reaches the minimum temperature, and output the global optimal solution as s * .

[0106] Specifically, the feasible solution construction method involved in steps 4.2.1 and 4.2.3 refers to determining the solution x of the transportation path and the solution y of the transportation volume, and then calculating the starting transportation time of each task according to the greedy strategy under the premise of meeting various transportation constraints, thereby constructing a complete feasible solution s=(x,y,t). The specific process is as follows:

[0107] Step 4.2.1.1, sort all tasks in order of priority from high to low while meeting the constraint conditions of the sequence relationship between tasks, and obtain a queue List;

[0108] Step 4.2.1.2, initialize network resource occupation information, the loading and unloading capacity of the station, the passing capacity of the road section, and the remaining capacity of other related transportation resources in each time window are set to the default maximum value;

[0109] Step 4.2.1.3, the first task of obtaining the first item of the List, calculates the available resources of the task according to the path involved in the task and the current network resource occupation, and sets the start transportation time of each path in the task according to the greedy strategy, that is, the task is executed as early as possible, and the current task is removed from the List after completion; repeat step 4.2.1.3 until the queue List is empty.

[0110] Specifically, the destruction operator involved in step 4.2.2 refers to deleting a part of the solution of the task from the feasible solution according to certain rules, and the repair operator refers to resetting the solution of the deleted task according to certain rules, and a pair of destruction-repair operators defines a neighborhood structure, and the following operators are used in the present application:

[0111] Destruction operator 1: randomly select 10% of the tasks and delete their solutions.

[0112] Destruction operator 2: randomly select 20% of the tasks and delete their solutions.

[0113] Destruction operator 3: select the task with the latest completion time and delete his solution.

[0114] Repair operator 1: for all deleted tasks, randomly reset several initial paths from the alternative path set, and randomly reset the transportation capacity of the path.

[0115] Repair operator 2: for all deleted tasks, keep the original transportation path unchanged, and randomly reset the transportation capacity of the path.

[0116] The above is only a preferred embodiment of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments, and any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that for ordinary technical personnel in the technical field, some improvements and refinements without departing from the principles of the present application shall be considered as the protection scope of the present application.

Claims

1. A method for optimizing the scheduling of emergency supplies transportation based on a multimodal transport network, characterized in that, include: Step 1: Establish a multimodal transport network for emergency supplies transportation. The multimodal transport network includes a hierarchical topology, station loading and unloading capacity, and road segment traffic capacity. Step 2: Input the set of emergency material transportation tasks within the planned time period. The task set includes the task start point, task end point, task transportation volume, earliest start time, latest delivery time, and task priority. Step 3: For the tasks in the multimodal transport network and emergency material transport task set, use the multi-objective shortest path algorithm to generate multimodal transport paths between the origin and destination of each task, and use the random perturbation method to change the weight coefficients of the optimization objectives to construct a set of alternative paths. Step 4: Based on the set of alternative routes, establish an objective function and its constraints with the goal of minimizing the overall transportation plan completion time. Use an adaptive large-scale neighborhood search algorithm to solve the objective function and obtain the optimal overall scheduling scheme.

2. The emergency material transportation optimization and scheduling method based on multimodal transport networks as described in claim 1, characterized in that, In step 1, the establishment of a multimodal transport network for emergency supplies specifically involves: Let the directed graph G = (V, E) represent the multimodal transport network, where the set of stations is V = {v0, v1, ..., v...}. n }, v n This represents the nth station; the set of road segments is E = {(v i v j ):v i v j ∈V, i≠j}, representing station v i With site v j Connection relationships between them; subgraph G f =(V f E f () represents the hierarchical transportation network for different modes of transport, f∈F:={road, rail, waterway, air}, V f and E f Let V represent the set of stations and the set of road segments in transportation mode f, respectively, and F represent the set of transportation modes, where V = V1∪V2∪...∪V |F| E = E1∪E2∪...∪E |F| ∪E t E t V represents the set consisting of all transfer routes. |F| E represents the set of stations for the last mode of transportation. |F| This represents the set of road segments that represent the last mode of transportation. For site v i ∈V, This indicates the site v i The maximum loading capacity per unit time. This indicates the site v i The maximum unloading capacity per unit time; For general road sections (v i v j )∈E\E t , l i，j c represents the travel distance of a road segment. i，j Indicates the travel time of a road segment, h i，j This indicates the maximum traffic capacity of a road segment within a unit of time. For the transfer route (v i v j )∈E t , This indicates the transfer rate.

3. The emergency material transportation optimization and scheduling method based on multimodal transport networks as described in claim 1, characterized in that, Step 3 specifically includes: Step 3.1: Select four indicators—route length, route time, number of transshipments, and transportation cost—as the optimization objectives for multimodal transport routes. Use the time value coefficient to normalize the optimization objectives across different dimensions, addressing the origin and destination points in each transport pair. The weight coefficients of the optimization objective are w1, w2, w3, and w4, respectively, satisfying ∑ i w i =1, i = 1, 2, 3, 4 and w i ≥0, randomly generate a set of weight coefficient combinations for the optimization objective. The optimal path under the current conditions is calculated using Dijkstra's algorithm on a directed graph and used as the multimodal transport path. Step 3.2: For the generated multimodal transport route, determine whether it meets the following criteria: whether the route length is less than the maximum allowed value, whether the route time is less than the maximum allowed value, and whether the number of transshipments is less than the maximum allowed value. If all criteria are met, the generated multimodal transport route is added to the candidate route set. Step 3.3: Apply random perturbation to the weight coefficient combination, and use Dijkstra's algorithm on the directed graph to calculate the optimal path under the current conditions to generate a new multimodal transport path; Step 3.4: Repeat steps 3.3.2 and 3.3.3, and delete duplicate paths in the alternative path set until the number of paths in the alternative path set meets the requirement.

4. The emergency material transportation optimization and scheduling method based on multimodal transport networks as described in claim 1, characterized in that, Step 4 includes: Step 4.1: Establish an objective function with the goal of minimizing the overall transportation plan's completion time. The formula is as follows: Where z represents the completion time of the overall transportation plan, and t is a real decision variable. r The start time of task r is represented by the decision variable θ, which is 0-1. p This indicates whether p from the set of candidate paths is selected as the transportation path for task r. The value is 1 if selected, and 0 otherwise. P(r) represents the set of candidate paths for task r. R represents the transit time for path p; R represents the total set of alternative paths. The constraints of the objective function are as follows: Constraint 1: Guarantee that each task is assigned at least one path, and the number of paths assigned to each task does not exceed P. num The bar, expressed as a formula, is: In the formula, θ p Represents a 0-1 decision variable; Constraint 2: In the formula, integer variable δ p q represents the transport volume carried by path p. r This represents the transportation requirements of task r; Constraint 3: Within each unit time window, the loading volume at any station shall not exceed the upper limit of the loading capacity, as expressed by the formula: In the formula, Indicates whether path p is at site v i Loading activity is performed, 0-1 variable μ p，k Indicates whether path p is loaded within time window k; V is the set of sites; This indicates the maximum loading capacity of the site per unit of time; K represents the overall planning time window. Constraint 4: Within each unit time window, the unloading volume of any site shall not exceed the upper limit of its unloading capacity, as expressed by the formula: In the formula, Indicates whether path p is at site v i To perform the uninstallation activity, the 0-1 variable v p，k Indicates whether path p is loaded within time window k; This indicates the site v i The maximum unloading capacity per unit time; Constraint 5: Within each unit time window, the traffic volume of any road segment shall not exceed the upper limit of its capacity, as expressed by the formula: In the formula, parameter b i，j，p This indicates that path p is a route segment (v) i v j ), 0-1 variable ξ p，k Indicates whether path p is transported within time window k; E is the set of road segments; Constraint 6: i p ,m p，k ,v p，k ,x p，k ∈{0,1} Constraint 7: t r ,d p ≥0 Step 4.2: Solve the objective function using the adaptive large-scale neighborhood search algorithm to obtain the optimal overall scheduling scheme.

5. The emergency material transportation optimization and scheduling method based on multimodal transport networks as described in claim 4, characterized in that, The specific steps for solving the objective function using the adaptive large-scale neighborhood search algorithm are as follows: Step 4.2.1: For each transportation task, randomly select several initial routes from the set of alternative routes. And randomly assign the transportation volume to the route. Get the start time of each task. The feasible solution construction method is called to obtain the initial solution s0 = (x0, y0, t0); x0 represents the initial solution of the transportation path, y0 represents the initial solution of the transportation volume, and t0 represents the initial solution of the transportation time. Step 4.2.2: Initialize the optimal solution s * Given the current solution s as the initial solution s0, set the initial temperature, final temperature, cooling rate and maximum number of iterations in the Metropolis acceptance criterion, and initialize the weights ρ of the destruction operator and the repair operator to the same value; Step 4.2.3: Based on the selection weights ρ of the destruction and repair operators, select a pair of destruction and repair operators from the neighborhood set according to the roulette wheel principle, perform a neighborhood search operation, and call the feasible solution construction method to generate a new solution. This represents the current solution for the transportation route. This represents the current solution for the transportation volume. The current solution represents the transit time; Step 4.2.4: Compare the new solutions If a new solution is better than the current solution s, then the new solution is used. Replace the current solution s; otherwise, determine whether to accept the new solution according to the Metropolis acceptance criterion, and update the historical best solution s. * ; Step 4.2.5: By collecting the performance of different destruction and repair operators in the past 100 iterations, score the destruction and repair operators and update the selection weights ρ of the destruction and repair operators. The scoring mechanism includes: (a) a new global optimal solution was obtained in the previous iteration; (b) a solution that was not found before was obtained in the previous iteration, but the objective function value of the solution is worse than the current solution; (c) a solution that was not found before was obtained in the previous iteration, but the objective function value of the solution is worse than the current solution, but the solution was accepted. Step 4.2.6: Repeat steps 4.2.3 to 4.2.5 until the maximum number of iterations is reached or the temperature reaches the minimum. Output the global optimal solution as s. * .

6. The emergency material transportation optimization and scheduling method based on multimodal transport networks as described in claim 5, characterized in that, The feasible solution construction method refers to, after determining the solutions for transportation routes and transportation quantities, calculating the start time of each task using a greedy strategy while satisfying various transportation constraints, thereby constructing a complete feasible solution. The specific process is as follows: Step 4.2.1.1 Sort all tasks in descending order of priority, while satisfying the constraints of the sequence relationship between tasks during sorting, to obtain a queue List; Step 4.2.1.2: Initialize network resource occupancy information. Set the loading and unloading capacity of stations, the traffic capacity of road segments, and the remaining capacity of other related transportation resources in each time window to the default maximum value. Step 4.2.1.3: Get the first task in the List, calculate the available resources for the task based on the path involved in the task and the current network resource usage, and set the start time of each path in the task according to the greedy strategy, that is, execute the task as early as possible. After completion, remove the current task from the List; repeat step 4.2.1.3 until the queue List is empty.

7. The emergency material transportation optimization and scheduling method based on multimodal transport networks as described in claim 5, characterized in that, The destruction operator refers to a solution that deletes a portion of the tasks from a feasible solution according to certain rules, and the repair operator refers to a solution that resets the deleted tasks according to certain rules. A pair of destruction-repair operators defines a neighborhood structure, and the destruction and repair operators are as follows: Destruction Operator 1: Randomly select 10% of the tasks and delete their solutions; Destruction Operator 2: Randomly select 20% of the tasks and delete their solutions; Destruction operator 3: Select the task with the latest completion time and delete its solution; Repair operator 1: For all deleted tasks, randomly reset several initial paths from the set of alternative paths, and randomly reset the transportation volume carried by the paths; Repair operator 2: For all deleted tasks, keep the original transportation path unchanged and randomly reset the transportation volume of the path.

Images