An emergency traffic collaborative decision method, system and device for a two-level scheduling system

Abstract

The application discloses an emergency traffic collaborative decision method and system for a two-level scheduling system and equipment, and belongs to the technical field of intelligent traffic, emergency logistics scheduling and traffic simulation. The method comprises the following steps: acquiring emergency event information, emergency demand information, local sub-center inventory information, multi-mode joint transportation network information, transportation resource information and constraint parameter information; constructing an emergency traffic collaborative decision model for a two-level scheduling system; and performing collaborative optimization on the emergency traffic collaborative decision model for a two-level scheduling system. The candidate collaborative decision scheme is introduced into an emergency traffic simulation deduction system to obtain corresponding simulation evaluation results. The target emergency traffic collaborative decision scheme is determined, and the plan selection results, execution sequence, departure time arrangement and transportation resource allocation results are output. The application improves the overall collaboration and unified scheduling capability in the emergency transportation process, and enhances the verifiability, explainability and executability of the emergency traffic decision scheme.

Description

Technical Field

[0001] This invention relates to an emergency traffic collaborative decision-making method, system, and equipment for a two-level dispatching system, belonging to the fields of intelligent transportation, emergency logistics dispatching, and traffic simulation technology. Background Technology

[0002] Existing emergency logistics research focuses on local route optimization, single transportation mode scheduling, or resource allocation under static conditions, which is insufficient to meet the needs of global collaborative decision-making under multi-center, multi-mode, and multi-constraint conditions.

[0003] As emergency scenarios become increasingly complex, existing emergency transportation decision-making methods suffer from several shortcomings. First, local sub-centers typically develop multiple candidate transportation plans based on local inventory and accessible routes, while the central government needs to coordinate the combination and execution order of these plans under global resource constraints, creating a significant coupling between the two. Second, emergency transportation often relies on multi-modal transport networks including highways, railways, waterways, and aviation. Differences in capacity limits, route capacity, time continuity, and priorities among different modes of transport result in high-dimensionality, strong constraints, and large-scale characteristics for the decision-making problem. Third, traditional heuristic or static rule-based methods suffer from low solution efficiency, insufficient adaptability to dynamic changes, and a lack of pre-execution verification mechanisms in complex emergency scenarios, making it difficult to provide efficient and reliable decision support for actual emergency command.

[0004] Meanwhile, existing emergency transportation solutions typically only focus on optimizing output results, lacking the ability to import candidate solutions into a simulation environment for dynamic simulation and verification. They also fail to identify key congestion points, resource conflict locations, and execution bottlenecks in advance. Therefore, how to construct an emergency transportation collaborative decision-making method for a two-level dispatching system, and how to dynamically evaluate candidate solutions using a simulation system, has become an urgent technical problem to be solved. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide an emergency traffic collaborative decision-making method, system and equipment for a two-level dispatch system.

[0006] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution.

[0007] In a first aspect, the present invention provides an emergency traffic collaborative decision-making method for a two-level dispatching system, comprising the following steps:

[0008] Acquire information on emergency events, emergency needs, local sub-center inventory, multimodal transport networks, transport resources, and constraint parameters;

[0009] Based on the emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, transportation resource information, and constraint parameter information, an emergency transportation collaborative decision-making model for a two-level dispatch system is constructed, and the emergency transportation collaborative decision-making model for a two-level dispatch system is collaboratively optimized under resource-constrained conditions.

[0010] The collaborative optimization of the emergency traffic collaborative decision-making model for a two-level dispatch system under resource-constrained conditions includes:

[0011] Each local sub-center determines the inventory materials, destination demand, feasible transportation modes, feasible transportation routes, and available transportation resources based on emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, transportation resource information, and constraint parameter information. Multiple feasible emergency transportation plans are generated based on the inventory materials, destination demand, feasible transportation modes, feasible transportation routes, and available transportation resources.

[0012] The central command center selects a combination of target plans from the feasible emergency transportation plans generated by various local sub-centers, and performs collaborative optimization on the execution order of the combination of target plans to obtain candidate collaborative decision-making schemes.

[0013] The candidate collaborative decision-making schemes are imported into a pre-built emergency traffic simulation system to dynamically simulate the road network resource occupation, transportation conflicts, task completion process and key node status under different schemes, and obtain the corresponding simulation evaluation results.

[0014] Based on the simulation evaluation results, a target emergency traffic collaborative decision-making scheme is determined, and the scheme selection results, execution order, departure time arrangement and transportation resource allocation results are output.

[0015] Furthermore, the emergency traffic collaborative decision-making model for the two-level dispatch system includes a local sub-center contingency plan decision-making layer and a central command center execution sequence decision-making layer;

[0016] At the local sub-center contingency plan decision-making level, each local sub-center generates its own corresponding contingency plan set, which includes candidate contingency plans under different transportation modes, different transportation routes and different resource allocation relationships;

[0017] At the central command center's execution order decision-making level, under the constraints of global resources, the central command center selects a combination of target plans from the plan sets of various local sub-centers, and generates an execution sequence for the combination of target plans according to preset execution order rules or optimization rules.

[0018] After the execution sequence is generated, the target plan combination and execution sequence are imported into the emergency traffic simulation system to obtain simulation evaluation results.

[0019] The selection results and execution order of the proposed plans are updated based on the simulation evaluation results to form a collaborative feedback adjustment mechanism.

[0020] Furthermore, the multimodal transport network is constructed as follows:

[0021] Define a set of available modes of transport, which includes at least road, rail, waterway and air transport;

[0022] A separate transportation network is constructed for each transportation mode, and the transportation network consists of a set of nodes and a set of edges;

[0023] The node set includes at least the station number, station name, geographical location, and mode of transportation attribute;

[0024] The edge set includes at least the route number, the origin and destination node numbers, the transportation mode attribute, and the route distance attribute;

[0025] A multi-mode joint transportation network is formed based on the set of nodes and edges under each transportation mode. Nodes and routes that are irrelevant to the current emergency scenario are removed from the emergency traffic simulation system to obtain a simulation network oriented towards the current emergency event.

[0026] Furthermore, the emergency traffic collaborative decision-making model for the two-level dispatch system takes minimizing the total task completion time and balanced utilization of transportation resources as its joint optimization objectives, which are expressed as:

[0027] ;

[0028] Where F is the collaborative optimization objective function, f1 is the total task completion time objective, f2 is the transportation resource balance utilization objective, β1 is the transportation timeliness weight parameter, β2 is the transportation balance weight parameter, and min means minimization.

[0029] Furthermore, the emergency traffic collaborative decision-making model for the two-level dispatch system includes network structure constraints, time-segment occupancy relationship constraints, segment capacity constraints, time continuity constraints, and material priority constraints.

[0030] The network structure constraints are expressed as follows:

[0031] ;

[0032] In the formula, I is the set of local subcenters, and L i Let x be the set of possible alternatives for local sub-center i. il Select variables for the contingency plan. This indicates the existence of logical symbols;

[0033] The time-segment occupancy relationship constraint is expressed as follows:

[0034] ;

[0035] In the formula, When selecting an optional contingency plan l for local subcenter i, the occupancy variable of road segment a under transportation mode m at time t, where M represents the set of transportation modes, and A m Let T represent the set of road segments under transportation mode m, and let T represent the set of discrete time periods within the planning time range;

[0036] The road segment capacity constraint is expressed as follows:

[0037] ;

[0038] In the formula, The occupancy variable of road segment a at time t is the optional contingency plan l. This represents the maximum transport capacity of segment a under transport mode m.

[0039] The time continuity constraint is expressed as:

[0040] ;

[0041] In the formula, Let j+1 be the time when the (j+1)th group of supplies from local sub-center i arrives at its destination. Let j be the time when the j-th group of supplies from local sub-center i arrives at its destination. The maximum allowable arrival time interval between material groups in local subcenter i;

[0042] The material priority constraint is expressed as follows:

[0043] ;

[0044] In the formula, and These represent the priorities of the resources corresponding to local sub-center i and local sub-center k, respectively. and These represent the execution sequence numbers of local subcenter i and local subcenter k, respectively.

[0045] This invention establishes a multi-modal joint transportation network and comprehensively considers constraints such as network structure, time-segment occupancy relationship, segment capacity, time continuity, and material priority. It achieves joint optimization of multi-center, multi-modal emergency transportation tasks under resource-constrained conditions and improves the model's adaptability to complex emergency scenarios.

[0046] Furthermore, the process of collaboratively optimizing the execution order of the target plan combination adopts a hybrid reinforcement learning solution mechanism, which includes the Q-Learning algorithm and the adaptive large-scale neighborhood search ALNS algorithm.

[0047] The ALNS algorithm is used to generate candidate state transition paths for the execution sequence of local subcenters;

[0048] The Q-Learning algorithm is used to iteratively learn and update the action selection of feasible plans under a given state transition path;

[0049] The hybrid reinforcement learning solution mechanism forms a joint solution framework for the two-level decision-making process by alternately optimizing the selection of plans and the execution order.

[0050] This invention improves the efficiency of plan combination selection and execution order optimization under large-scale complex constraints by introducing a hybrid solution mechanism of Q-Learning and ALNS. Experiments show that as the number of transportation plans increases, the QL-ALNS hybrid algorithm exhibits better efficiency and faster convergence speed than either Q-Learning or ALNS alone. When the number of plans is 90, its solution time is 61 seconds, which is about 50% faster than Q-Learning alone (123 seconds) and about 22% faster than ALNS alone (78 seconds). In terms of convergence, QL-ALNS converges in about 80 iterations, while ALNS requires about 120 iterations and Q-Learning requires about 240 iterations.

[0051] Furthermore, the process of the hybrid reinforcement learning solution mechanism includes:

[0052] In the action selection process of the Q-Learning algorithm, an action masking mechanism is set to block actions that do not meet the constraints, and only executable actions are retained to participate in the policy update.

[0053] Construct a dual-objective reward mechanism that simultaneously reflects transportation timeliness and resource balance, so as to achieve synchronous optimization of task completion time and resource utilization balance;

[0054] In the ALNS algorithm, a disruption operator strategy based on dynamic bottleneck removal is set to identify and remove execution sequence segments that affect transportation efficiency.

[0055] In the ALNS algorithm, a repair operator strategy based on temporal compact interpolation is set to re-insert and reconstruct the removed execution order fragments to obtain new candidate solutions for the execution order.

[0056] Furthermore, the emergency traffic simulation system includes a road network construction phase, a scheme import phase, a dynamic simulation phase, and a result evaluation phase.

[0057] The road network construction phase is used to build an emergency traffic simulation network based on a multimodal joint transportation network.

[0058] The scheme import phase is used to import the feasible emergency transportation plan generated by the local sub-center and the execution sequence generated by the central command center into the simulation environment.

[0059] The dynamic simulation phase is used to simulate node occupancy, road segment resource occupancy, material departure and arrival processes, transportation resource conflicts, and the evolution of key congestion nodes under different transportation modes.

[0060] The results evaluation phase is used to output the total task completion time, resource utilization balance, congestion at key nodes, transportation conflicts, and collaborative decision-making schemes.

[0061] Secondly, the present invention provides an emergency traffic collaborative decision-making system for a two-level dispatching system, comprising:

[0062] The data acquisition module is used to acquire emergency event information, emergency demand information, local sub-center inventory information, multimodal transport network information, transport resource information, and constraint parameter information.

[0063] The contingency plan generation module is used by local sub-centers to determine inventory materials, destination demand, feasible transportation modes, feasible transportation routes and available transportation resources based on emergency event information, emergency demand information, local sub-center inventory information, multi-mode joint transportation network information, transportation resource information and constraint parameter information. Based on the inventory materials, destination demand, feasible transportation modes, feasible transportation routes and available transportation resources, multiple feasible emergency transportation contingency plans are generated.

[0064] The collaborative modeling module is used to build an emergency traffic collaborative decision-making model for a two-level dispatch system;

[0065] The sequence optimization module is used by the central command center to select a combination of target plans from the feasible emergency transportation plans generated by various local sub-centers, and to perform collaborative optimization on the execution order of the combination of target plans to obtain candidate collaborative decision schemes.

[0066] The simulation and deduction module is used to import the candidate collaborative decision-making schemes into a pre-built emergency traffic simulation and deduction system, and to dynamically deduce the road network resource occupation, transportation conflicts, task completion process and key node status under different schemes, and obtain the corresponding simulation evaluation results.

[0067] The decision output module is used to determine the target emergency traffic collaborative decision-making scheme based on the simulation evaluation results, and output the scheme selection results, execution order, departure time arrangement and transportation resource allocation results.

[0068] Thirdly, the present invention provides a computer device, comprising,

[0069] One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing the method described in the first aspect.

[0070] The beneficial effects achieved by this invention are as follows:

[0071] This invention solves the problem of the disconnect between central coordination and local execution in existing emergency transportation organization by constructing a two-level dispatch architecture that allows the central command center and local sub-centers to work together, thereby improving the overall coordination and unified dispatch capabilities in the emergency transportation process.

[0072] This invention constructs an emergency traffic simulation environment, imports candidate collaborative decision-making schemes into the simulation environment, and dynamically evaluates road network occupancy, resource conflicts, task completion processes, and key congestion nodes during scheme execution, thereby enhancing the verifiability, interpretability, and executability of emergency traffic decision-making schemes. In this embodiment, the simulation system includes four stages: road network construction, scheme import, dynamic simulation, and result evaluation. Attached Figure Description

[0073] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0074] Figure 2 This is a schematic diagram of the emergency traffic collaborative decision-making framework for a two-level dispatch system according to the present invention;

[0075] Figure 3 This is a schematic diagram illustrating the construction of the multimodal joint transportation network of the present invention;

[0076] Figure 4 This is a schematic diagram illustrating the optimized process of local sub-center contingency plan generation and central command center execution order in this invention.

[0077] Figure 5 This is a schematic diagram of the hybrid reinforcement learning solution mechanism of this invention;

[0078] Figure 6 This is a schematic diagram of the emergency traffic simulation and deduction system of the present invention;

[0079] Figure 7 This is a schematic diagram of the simulated road network for the multi-mode emergency transportation system in an embodiment of the present invention;

[0080] Figure 8 This is a schematic diagram illustrating resource conflict and critical node analysis in an embodiment of the present invention;

[0081] Figure 9 This is a schematic diagram comparing task completion time and resource utilization balance in an embodiment of the present invention;

[0082] Figure 10 This is a schematic diagram of the emergency traffic collaborative decision-making simulation system for a two-level dispatching system, as described in this invention. Detailed Implementation

[0083] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0084] Example 1, as Figure 1 As shown in the figure, this embodiment introduces an emergency traffic collaborative decision-making method for a two-level dispatch system. The method includes the following steps:

[0085] S101: Acquire emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, transportation resource information, and constraint parameter information to establish an emergency transportation collaborative decision-making model for a two-level dispatch system;

[0086] S102: Based on a hybrid reinforcement learning solution mechanism, feasible emergency transportation plans are combined and the execution order is optimized to obtain candidate collaborative decision-making schemes;

[0087] S103: Input the candidate collaborative decision-making scheme into the pre-established emergency traffic simulation environment for dynamic simulation processing, obtain the simulation evaluation results, and output the target emergency traffic collaborative decision-making scheme based on the simulation evaluation results.

[0088] The following provides further explanation of each step.

[0089] Before implementation, this embodiment requires scenario data input. Specifically, a regional emergency transportation system comprising a central command center, multiple local sub-centers, multiple demand points, and various transportation modes is selected as the decision-making object. The central command center is responsible for formulating a global transportation plan combination in both time and space dimensions, while local sub-centers are responsible for generating multiple alternative transportation plans and managing specific transportation processes. This problem is essentially a two-level scheduling decision-making problem oriented towards the collaborative operation of the central command center and local sub-centers: the first-level decision-making layer is the contingency plan decision for local sub-centers, and the second-level decision-making layer is the execution sequence decision for the central command center. The two-level decision-making processes influence each other and together constitute a high-dimensional combinatorial optimization problem.

[0090] like Figure 2 and Figure 3As shown, the specific process of S101 is as follows:

[0091] S11. Obtain emergency event information, emergency demand information, local sub-center inventory information, multimodal transport network information, and transport resource information;

[0092] Specifically, emergency event information should include at least the event type, location, time, and scope of impact; emergency demand information should include at least the location of the demand point, the type and quantity of supplies, and the urgency level of the supplies; local sub-center inventory information should include at least the type and quantity of inventory supplies and the earliest departure time; and transportation resource information should include at least the available capacity limit, available time window, and route resources under different transportation modes. These data collectively constitute the scenario input for emergency transportation collaborative decision-making.

[0093] S12. Construct a multimodal transport network.

[0094] Let the set of available transportation modes be . , in These correspond to road, rail, waterway, and air transport, respectively; for each mode of transport... Construct transportation networks separately:

[0095] ;

[0096] in, Representation pattern The set of nodes below, Representation pattern The set of edges is defined below; and a unified multimodal joint transportation network is formed through pattern coupling relationships.

[0097] In this embodiment, the location of the emergency event is set to a certain city, and at least 30 cities are set to respond to the emergency event. The transportation modes include road, rail, waterway and air. The multimodal transport network data comes from OpenStreetMap (OSM) and China Railway Map, with a total of 668 station data and 22,919 route data.

[0098] S13. Establish a local sub-center contingency plan decision-making level.

[0099] Let the local sub-center set be defined. For each local sub-center A set of optional contingency plans is generated in advance. The optional contingency plans include at least candidate solutions under different combinations of transportation modes, different transportation routes, and different resource allocation relationships. Each local sub-center must and can only select one contingency plan, calculated as follows:

[0100] ;

[0101] In the formula, Selecting variables for a binary contingency plan, where the local sub-center Select contingency plan hour, ,otherwise, .

[0102] S14. Establish a central command center to implement the decision-making process.

[0103] Under the constraints of overall resources, the central command center combines and selects candidate transportation plans submitted by various local sub-centers, and determines the execution order of the selected plans by each local sub-center. Let... Indicates local sub-center If the execution sequence number is specified, then the decision-making outcome at the central level includes not only the selected combination of contingency plans but also the execution sequence. .

[0104] S15. Construct a collaborative optimization objective function under resource-constrained conditions.

[0105] In this embodiment, the two-level emergency transportation collaborative decision-making model takes minimizing the total task completion time and balanced utilization of transportation resources as its joint optimization objectives, and its calculation formula is as follows:

[0106] ;

[0107] In the formula, To collaboratively optimize the objective function, The overall task completion time target. To achieve the goal of balanced utilization of transportation resources, For the weighting parameter of transportation timeliness, This is the weighting parameter for transportation balance.

[0108] In this embodiment, the total task completion time target can be expressed as:

[0109] ;

[0110] In the formula, Indicates local sub-center The delivery completion time.

[0111] The resource usage of transportation mode m can be expressed as:

[0112] ;

[0113] In the formula, Indicates local sub-center Select contingency plan Time-based transportation modes Resource usage.

[0114] ;

[0115] in, This represents the average resource usage across all transportation modes.

[0116] S16. Construct constraints.

[0117] In this embodiment, the two-level emergency traffic collaborative decision-making model comprehensively considers network structure constraints, time-segment occupancy relationship constraints, segment capacity constraints, time continuity constraints, and material priority constraints.

[0118] The network structure constraints are expressed as follows:

[0119] ;

[0120] In the formula, Select variables for the contingency plan. Let i be the set of possible alternative plans for local subcenter i.

[0121] The time-segment occupancy relationship constraint is expressed as follows:

[0122] ;

[0123] In the formula, When selecting plan l for local subcenter i, the occupancy variable of road segment a under transportation mode m at time t.

[0124] The road segment capacity constraint is used to limit the total resources consumed by each transportation task on any road segment at any time, under any transportation mode, to not exceed the maximum capacity of the corresponding road segment. Its calculation formula is:

[0125] ;

[0126] In the formula, Indicates local sub-center contingency plan In time For the section The occupancy variable, Represents a set of transportation modes. Indicates transportation mode The following road sections are collected. This represents the set of time periods within the planning timeframe. Indicates transportation mode Lower section The maximum transport capacity.

[0127] The time continuity constraint is used to limit the maximum time interval between different batches of materials arriving at the destination from the same local subcenter. Its calculation formula is:

[0128] ;

[0129] In the formula, For local sub-centers The The time it takes for the supplies to arrive at their destination. For local sub-centers The maximum allowable time interval between material shipments;

[0130] The material priority constraint is used to ensure that transportation plans corresponding to high-priority materials are executed first. The material priority constraint is expressed as follows:

[0131] ;

[0132] In the formula, and These represent the priorities of the resources corresponding to local sub-center i and local sub-center k, respectively. and These represent the execution sequence numbers of local subcenter i and local subcenter k, respectively.

[0133] Material priority constraints are used to ensure that transportation plans corresponding to high-priority materials are executed first. Specifically, if the local sub-center... The priority of supplies is higher than that of local sub-centers. Then the local sub-center The corresponding execution sequence number is not greater than that of the local sub-center. The corresponding execution sequence number.

[0134] like Figure 4 and Figure 5 As shown, the specific process of S102 is as follows:

[0135] S21. Initialize the local sub-center execution sequence.

[0136] Since this problem involves two coupled decision-making processes—local sub-center selection and central execution order arrangement—it is difficult to naturally form a stable state transition path using only a single reinforcement learning process. Therefore, in this embodiment, the adaptive large-scale neighborhood search (ALNS) algorithm is first used to initialize the local sub-center execution order, generating an initial execution sequence. .

[0137] S22. Establish the Q-Learning state space and action space.

[0138] In this embodiment, the current local sub-center to be decided is regarded as the state. The optional transportation plan corresponding to the local subcenter is regarded as an action. For any given state, the action space consists of a set of candidate plans for the corresponding local sub-center. To avoid violating resource constraints, priority constraints, and time continuity constraints, an action masking mechanism is introduced during the action selection process, retaining only actions that satisfy the constraints into the set of executable actions.

[0139] S23. Selecting actions according to the contingency plan and updating the Q table.

[0140] In a given execution sequence In this context, Q-Learning, as a model-free reinforcement learning method, sequentially selects plans for each local sub-center according to the execution sequence. After each action is executed, the reward value is calculated based on the arrival time of the last batch of supplies at the destination and the resource utilization balance, and the state-action value is updated using the standard Q-value iteration formula:

[0141] ;

[0142] In the formula, Indicates the first After the iteration, in state Next action State—Action Value; Indicates the first State at the next iteration With action The corresponding state—action value; Indicates the first The next state during the next iteration Next, select candidate actions State-action value; This indicates the current status of the local sub-centers awaiting decision-making. Indicates the state The selected transportation plan actions; Indicates the execution of an action The next state after that; Indicates the next state Candidate actions from the set of executable actions; Indicates the execution of an action The instant reward obtained afterward; This represents the learning rate parameter; Indicates the discount factor; This represents the set of executable actions in the next state.

[0143] In this embodiment, the total reward value R adopts a bi-objective reward mechanism, and its calculation formula is as follows:

[0144] ;

[0145] In the formula, This indicates a reward for timely delivery. This indicates a reward for balanced resource utilization. The shorter the transportation completion time, the higher the reward. The higher the level, the more balanced the resource utilization. The higher.

[0146] S24. Disturb and reconstruct the execution order.

[0147] After completing one round of Q-Learning solution selection, ALNS (Adaptive Large Neighborhood Search) perturbs and reconstructs the current execution sequence. Specifically, a disruption operator strategy based on dynamic bottleneck removal is adopted to identify and remove execution sequence segments that affect transportation efficiency; then, a repair operator strategy based on temporal compact interpolation is used to re-insert and reconstruct the removed segments, thereby obtaining new candidate solutions for the execution sequence.

[0148] S25. Repeat the iteration until the stopping condition is met.

[0149] The reconstructed execution sequence is re-input into the Q-Learning module, and S22 to S24 are executed repeatedly until the preset number of iterations is reached or the objective function converges, and candidate collaborative decision schemes are output.

[0150] like Figure 6 As shown, the specific process of S103 is as follows:

[0151] S31. Construct an emergency traffic simulation environment.

[0152] In this embodiment, an emergency traffic simulation environment is constructed using a Python simulation program combined with the NetworkX road network modeling library. A simulation network oriented towards the current emergency scenario is built based on a multimodal transport network. Nodes and routes irrelevant to the current emergency task are removed, while key hubs, key road sections, and mode-switching nodes are retained to form the simulated road network in this embodiment. Figure 7 As shown, a regional multimodal emergency transportation simulation road network, comprising multiple city nodes, transportation routes, and key corridors, is formed based on a multimodal joint transportation network to support the dynamic simulation of subsequent candidate collaborative decision-making schemes.

[0153] S32. Import candidate collaborative decision-making schemes.

[0154] The results of the local sub-centers' contingency plan selection, the execution order of the central command center, the departure time arrangements for each transportation mode, and the resource allocation results are imported into the simulation environment.

[0155] S33, Execute dynamic simulation.

[0156] During the simulation, the process of dispatching, occupying, transferring, transiting, and arriving of materials from various local sub-centers in a multi-modal joint transportation network is simulated according to the simulation clock. Simultaneously, the resource occupancy of road segments, node congestion levels, resource conflict locations, key node status, and task completion status are recorded. For example... Figure 8 As shown, during the dynamic simulation process, resource conflict locations, key congestion nodes, and bottleneck sections in the road network can be identified and visualized, thus providing a basis for subsequent scheme evaluation and adjustment.

[0157] S34. Output simulation evaluation results.

[0158] After the simulation, the output includes the total task completion time, resource utilization balance, key congestion nodes, transportation conflicts, and the comprehensive execution effect of candidate collaborative decision-making schemes. Based on this, the target emergency traffic collaborative decision-making scheme is determined. Figure 9 As shown, the simulation evaluation results can be further presented in the form of resource utilization rate of different transportation modes, overall resource utilization balance, and optimal transportation completion time, which can be used to evaluate the execution effect of candidate collaborative decision-making schemes.

[0159] Specifically, the present invention will be further illustrated below through embodiments:

[0160] To verify the effectiveness of the hybrid reinforcement learning solution mechanism described in this invention in large-scale scenario scenarios, the method of this invention was compared with the standalone Q-Learning method and the standalone ALNS method under the same scenario parameters. The results are shown in Table 1.

[0161] Table 1. Comparison of efficiency and convergence of different algorithms in large-scale contingency planning scenarios.

[0162] Q-Learning 90 123 240 ALNS 90 78 120 QL-ALNS 90 61 80

[0163] As shown in Table 1, in a scenario with a total of 90 contingency plans, the QL-ALNS hybrid algorithm used in this invention outperforms the individual Q-Learning and ALNS algorithms in both solution time and convergence speed.

[0164] To further illustrate the feasibility and effectiveness of the emergency transportation collaborative decision-making method for a two-level dispatch system described in this invention under resource-constrained conditions, this embodiment uses a regional multi-mode emergency transportation system comprising a central command center, multiple local sub-centers, and four modes of transportation as the test scenario; the location of the emergency event is set as a certain city so that waterway and air transportation can also be reflected in the transportation plan; at least 30 cities are set to issue emergency material transportation plans in response to the emergency event to ensure that there are usage conflicts in the road network resources and that the solution space is of moderate size.

[0165] Example 2, based on the same inventive concept as Example 1, introduces an emergency traffic collaborative decision-making system for a two-level dispatching system, such as... Figure 10 As shown, it includes: a data acquisition module 11, a contingency plan generation module 12, a collaborative modeling module 13, a sequence optimization module 14, a simulation and deduction module 15, and a decision output module 16.

[0166] Data acquisition module 11 is used to acquire emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, transportation resource information, and constraint parameter information;

[0167] The contingency plan generation module 12 is used by each local sub-center to determine the inventory materials, destination demand, feasible transportation modes, feasible transportation routes and available transportation resources based on emergency event information, emergency demand information, local sub-center inventory information, multi-mode joint transportation network information, transportation resource information and constraint parameter information. Based on the inventory materials, destination demand, feasible transportation modes, feasible transportation routes and available transportation resources, multiple feasible emergency transportation contingency plans are generated.

[0168] Collaborative modeling module 13 is used to construct an emergency traffic collaborative decision-making model for a two-level dispatching system;

[0169] The sequence optimization module 14 is used by the central command center to select a combination of target plans from the feasible emergency transportation plans generated by various local sub-centers, and to perform collaborative optimization on the execution order of the combination of target plans to obtain candidate collaborative decision schemes.

[0170] The simulation and deduction module 15 is used to import the candidate collaborative decision-making schemes into the pre-built emergency traffic simulation and deduction system, and to dynamically deduce the road network resource occupation, transportation conflicts, task completion process and key node status under different schemes, and obtain the corresponding simulation evaluation results.

[0171] The decision output module 16 is used to determine the target emergency traffic collaborative decision-making scheme based on the simulation evaluation results, and output the scheme selection results, execution order, departure time arrangement and transportation resource allocation results.

[0172] Example 3, based on the same inventive concept as other examples, introduces a computer device including a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the steps of the above-described emergency traffic collaborative decision-making method for a two-level dispatch system. The steps include at least: acquiring emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, and transportation resource information; constructing an emergency traffic collaborative decision-making model for a two-level dispatch system; selecting and optimizing the execution order of feasible emergency transportation plans; and simulating and deducing candidate collaborative decision-making schemes and outputting a target emergency traffic collaborative decision-making scheme.

[0173] Example 4, based on the same inventive concept as other examples, introduces a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described emergency traffic collaborative decision-making method for a two-level dispatch system.

[0174] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0175] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0176] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0177] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0178] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An emergency traffic collaborative decision-making method for a two-level dispatch system, characterized in that, Includes the following steps: Acquire information on emergency events, emergency needs, local sub-center inventory, multimodal transport networks, transport resources, and constraint parameters; Based on the emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, transportation resource information, and constraint parameter information, an emergency transportation collaborative decision-making model for a two-level dispatch system is constructed, and the emergency transportation collaborative decision-making model for a two-level dispatch system is collaboratively optimized under resource-constrained conditions. The collaborative optimization of the emergency traffic collaborative decision-making model for a two-level dispatch system under resource-constrained conditions includes: Each local sub-center determines the inventory materials, destination demand, feasible transportation modes, feasible transportation routes, and available transportation resources based on emergency event information, emergency demand information, local sub-center inventory information, multi-modal joint transportation network information, transportation resource information, and constraint parameter information. Multiple feasible emergency transportation plans are generated based on the inventory materials, destination demand, feasible transportation modes, feasible transportation routes, and available transportation resources. The central command center selects a combination of target plans from the feasible emergency transportation plans generated by various local sub-centers, and performs collaborative optimization on the execution order of the combination of target plans to obtain candidate collaborative decision-making schemes. The candidate collaborative decision-making schemes are imported into a pre-built emergency traffic simulation system to dynamically simulate the road network resource occupation, transportation conflicts, task completion process and key node status under different schemes, and obtain the corresponding simulation evaluation results. Based on the simulation evaluation results, a target emergency traffic collaborative decision-making scheme is determined, and the scheme selection results, execution order, departure time arrangement and transportation resource allocation results are output. The emergency traffic collaborative decision-making model for the two-level dispatch system includes network structure constraints, time-segment occupancy relationship constraints, segment capacity constraints, time continuity constraints, and material priority constraints. The network structure constraints are expressed as follows: ； In the formula, I is the set of local subcenters, and L i Let x be the set of possible alternatives for local sub-center i. il Select variables for the contingency plan. This indicates the existence of logical symbols; The time-segment occupancy relationship constraint is expressed as follows: ； In the formula, This represents the occupancy variable of road segment a under transport mode m at time t when local subcenter i selects optional plan l, where M represents the set of transport modes, and A represents the occupancy variable of road segment a under transport mode m at time t. m Let T represent the set of road segments under transportation mode m, and let T represent the set of discrete time periods within the planning time range; The road segment capacity constraint is expressed as follows: ； In the formula, The occupancy variable of road segment a at time t is the optional contingency plan l. This represents the maximum transport capacity of segment a under transport mode m. The time continuity constraint is expressed as: ； In the formula, Let j+1 be the time when the (j+1)th group of supplies from local sub-center i arrives at its destination. Let j be the time when the j-th group of supplies from local sub-center i arrives at its destination. The maximum allowable arrival time interval between material groups in local subcenter i; The material priority constraint is expressed as follows: ； In the formula, and These represent the priorities of the resources corresponding to local sub-center i and local sub-center k, respectively. and These represent the execution sequence numbers of local subcenter i and local subcenter k, respectively.

2. The emergency traffic collaborative decision-making method for a two-level dispatch system according to claim 1, characterized in that, The emergency traffic collaborative decision-making model for a two-level dispatch system includes a local sub-center contingency plan decision-making layer and a central command center execution sequence decision-making layer. At the local sub-center contingency plan decision-making level, each local sub-center generates its own corresponding contingency plan set, which includes candidate contingency plans under different transportation modes, different transportation routes and different resource allocation relationships; At the central command center's execution order decision level, under the constraints of global resources, the central command center selects a combination of target plans from the plan sets of various local sub-centers, and generates an execution sequence for the combination of target plans according to preset execution order rules or optimization rules. After the execution sequence is generated, the target plan combination and execution sequence are imported into the emergency traffic simulation and deduction system to obtain simulation evaluation results; The selection results and execution order of the proposed plans are updated based on the simulation evaluation results to form a collaborative feedback adjustment mechanism.

3. The emergency traffic collaborative decision-making method for a two-level dispatch system according to claim 1, characterized in that, The construction of the multimodal transport network includes: Define the set of available modes of transport, which includes road, rail, waterway and air transport; A separate transportation network is constructed for each transportation mode, and the transportation network consists of a set of nodes and a set of edges; The node set includes station number, station name, geographical location, and transportation mode attribute; The edge set includes the route number, origin and destination node numbers, transportation mode attribute, and route distance attribute; A multi-mode joint transportation network is formed based on the set of nodes and edges under each transportation mode. Nodes and routes that are irrelevant to the current emergency scenario are removed from the emergency traffic simulation system to obtain a simulation network oriented towards the current emergency event.

4. The emergency traffic collaborative decision-making method for a two-level dispatch system according to claim 1, characterized in that, The emergency traffic collaborative decision-making model for a two-level dispatch system takes minimizing the total task completion time and balanced utilization of transportation resources as its joint optimization objectives. These joint optimization objectives are expressed as follows: ； In the formula, F is the collaborative optimization objective function, f1 is the total task completion time target, f2 is the transportation resource balanced utilization target, β1 is the transportation timeliness weight parameter, β2 is the transportation balance weight parameter, and min means minimization.

5. The emergency traffic collaborative decision-making method for a two-level dispatch system according to claim 1, characterized in that, The process of collaboratively optimizing the execution order of the target plan combination adopts a hybrid reinforcement learning solution mechanism, which includes the Q-Learning algorithm and the adaptive large-scale neighborhood search ALNS algorithm. The ALNS algorithm is used to generate candidate state transition paths for the execution sequence of local subcenters; The Q-Learning algorithm is used to iteratively learn and update the action selection of feasible plans under a given state transition path; The hybrid reinforcement learning solution mechanism forms a joint solution framework for the two-level decision-making process by alternately optimizing the selection of plans and the execution order.

6. The emergency traffic collaborative decision-making method for a two-level dispatch system according to claim 5, characterized in that, The process of the hybrid reinforcement learning solution mechanism includes: In the action selection process of the Q-Learning algorithm, an action masking mechanism is set to block actions that do not meet the constraints, and only executable actions are retained to participate in the policy update. Construct a dual-objective reward mechanism that simultaneously reflects transportation timeliness and resource balance, so as to achieve synchronous optimization of task completion time and resource utilization balance; In the ALNS algorithm, a disruption operator strategy based on dynamic bottleneck removal is set to identify and remove execution sequence segments that affect transportation efficiency. In the ALNS algorithm, a repair operator strategy based on temporal compact interpolation is set to re-insert and reconstruct the removed execution order fragments to obtain new candidate solutions for the execution order.

7. The emergency traffic collaborative decision-making method for a two-level dispatch system according to claim 5, characterized in that, The emergency traffic simulation system includes the road network construction stage, the scheme import stage, the dynamic simulation stage, and the result evaluation stage. The road network construction phase is used to build an emergency traffic simulation network based on a multimodal joint transportation network. The scheme import phase is used to import the feasible emergency transportation plan generated by the local sub-center and the execution sequence generated by the central command center into the simulation environment. The dynamic simulation phase is used to simulate node occupancy, road segment resource occupancy, material departure and arrival processes, transportation resource conflicts, and the evolution of key congestion nodes under different transportation modes. The results evaluation phase is used to output the total task completion time, resource utilization balance, congestion at key nodes, transportation conflicts, and collaborative decision-making schemes.

8. A system based on the emergency traffic collaborative decision-making method for a two-level dispatching system according to any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire emergency event information, emergency demand information, local sub-center inventory information, multimodal transport network information, transport resource information, and constraint parameter information. The contingency plan generation module is used by local sub-centers to determine inventory materials, destination demand, feasible transportation modes, feasible transportation routes and available transportation resources based on emergency event information, emergency demand information, local sub-center inventory information, multi-mode joint transportation network information, transportation resource information and constraint parameter information. Based on the inventory materials, destination demand, feasible transportation modes, feasible transportation routes and available transportation resources, multiple feasible emergency transportation contingency plans are generated. The collaborative modeling module is used to build an emergency traffic collaborative decision-making model for a two-level dispatch system; The sequence optimization module is used by the central command center to select a combination of target plans from the feasible emergency transportation plans generated by various local sub-centers, and to perform collaborative optimization on the execution order of the combination of target plans to obtain candidate collaborative decision schemes. The simulation and deduction module is used to import the candidate collaborative decision-making schemes into a pre-built emergency traffic simulation and deduction system, and to dynamically deduce the road network resource occupation, transportation conflicts, task completion process and key node status under different schemes, and obtain the corresponding simulation evaluation results. The decision output module is used to determine the target emergency traffic collaborative decision-making scheme based on the simulation evaluation results, and output the scheme selection results, execution order, departure time arrangement and transportation resource allocation results.

9. A computer device, characterized in that, include, One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing the method of any of claims 1 to 7.

Images