Search

CN-122022623-A - Multi-agent-based multi-intermodal dynamic path planning system and method

CN122022623ACN 122022623 ACN122022623 ACN 122022623ACN-122022623-A

Abstract

The application provides a multi-agent-based multi-intermodal dynamic path planning system and method. The system comprises a network construction module, a multi-agent collaborative risk processing module, a risk mapping module, a combination optimization solving module, a dynamic re-planning control module and a result output module, wherein the network construction module is used for constructing a multi-type intermodal transportation network diagram and generating a candidate path set, the multi-agent collaborative risk processing module is used for driving a plurality of agents to execute risk information acquisition and risk parameter generation processing on a road segment task set in parallel, the risk mapping module is used for updating side weights associated with corresponding transportation road segments in the multi-type intermodal transportation network diagram based on road segment risk parameters to obtain a risk weighted transportation network diagram, the combination optimization solving module is used for outputting a transportation path scheme corresponding to a demand parameter, the dynamic re-planning control module is used for updating road segment risk parameters for the affected transportation road segment set, and the result output module is used for outputting the transportation path scheme and the risk information. The application can improve risk perception and quantization efficiency, promote path dynamic update speed, and enhance robustness and interpretability of planning results.

Inventors

  • WU QINGYAO

Assignees

  • 厦门屿智未来科技有限公司

Dates

Publication Date
20260512
Application Date
20251230

Claims (10)

  1. 1. A multi-agent based multi-modal dynamic path planning system, comprising: the network construction module is used for receiving a starting point, an ending point and demand parameters associated with transportation constraint conditions, constructing a multi-type intermodal transportation network diagram and generating a candidate path set based on the multi-type intermodal transportation network diagram; the multi-agent cooperative risk processing module is used for mapping the transportation road segments associated with the candidate path set into a road segment task set, and driving a plurality of agents to execute risk information acquisition and risk parameter generation processing on the road segment task set in parallel so as to obtain road segment risk parameters associated with each transportation road segment; the risk mapping module is used for updating the side weights associated with the corresponding transportation road sections in the multi-road intermodal transportation network diagram based on the road section risk parameters to obtain a risk weighted transportation network diagram; The combined optimization solving module is used for constructing a combined optimization model meeting the transportation constraint conditions on the risk weighted transportation network diagram, solving the combined optimization model and outputting a transportation path scheme corresponding to the requirement parameters; The dynamic re-planning control module is used for determining an affected transportation road section set when external risk information changes, driving the multi-agent cooperative risk processing module to update road section risk parameters for the affected transportation road section set, driving the risk mapping module to update side weights and driving the combined optimization solving module to solve again so as to update the transportation path scheme; And the result output module is used for outputting the transportation path scheme and risk information associated with the transportation path scheme.
  2. 2. The system of claim 1, wherein the receiving the start point, the end point, and the demand parameters associated with the transportation constraint, constructing a multi-modal transportation network map, comprises: determining a node set and an edge set of the multi-mode intermodal transportation network diagram based on the demand parameters, wherein the node set is used for representing a starting point, an ending point and at least one transit hub, and the edge set is used for representing feasible transportation road sections adopting different transportation modes among the nodes; And writing basic transportation parameters associated with corresponding transportation road segments for each side, and performing feasibility screening on the multi-mode intermodal transportation network map based on the basic transportation parameters and the transportation constraint conditions to obtain a feasible subgraph for generating a candidate path set.
  3. 3. The system of claim 2, wherein the generating a set of candidate paths based on the multi-modal intermodal transportation network map comprises: Performing a path search process on the viable subgraph based on a start point and an end point to obtain at least one initial path formed by sequentially connecting a plurality of transportation segments; Calculating path parameters corresponding to the transportation constraint conditions for the initial paths based on the basic transportation parameters corresponding to the transportation road sections, and executing path screening and path pruning processing for the initial paths according to the transportation constraint conditions to obtain candidate path sets meeting the transportation constraint conditions; each candidate route in the set of candidate routes is normalized to a route structure including a sequence of transportation segments, each transportation segment in the route structure being associated with a corresponding edge in the multi-modal intermodal transportation network map.
  4. 4. The system of claim 1, wherein the mapping the transportation segments associated with the candidate set of paths to a set of segment tasks comprises: Extracting transport road segments corresponding to each candidate path based on the candidate path set, and determining associated road segment attribute information for each transport road segment; For each transportation road section, generating a risk investigation task list associated with the transportation road section based on the road section attribute information, wherein the risk investigation task list comprises a risk information acquisition task and a risk parameter generation task corresponding to the risk information acquisition task; And carrying out priority ordering and dependency relation arrangement on each risk investigation task list based on the road section attribute information and a preset task arrangement rule to form a road section task set for driving a plurality of agents to execute in parallel.
  5. 5. The system of claim 4, wherein the driving the plurality of agents to perform the risk information acquisition and risk parameter generation process on the set of road segment tasks in parallel to obtain road segment risk parameters associated with each transportation road segment comprises: Distributing each risk information acquisition task in the road section task set to a first type of agent for execution, so as to acquire external risk information from a multi-source data source aiming at a corresponding transportation road section, and executing deduplication, correlation screening and credibility evaluation on the external risk information to obtain a risk event set associated with the transportation road section; distributing each risk parameter generation task in the road section task set to a second class of agent for execution, so as to carry out consistency check and numerical processing on the risk event set and generate road section risk parameters associated with the transportation road section; In the parallel execution process, multiplexing and updating of the risk event set and the road section risk parameters among different transportation road sections are achieved through sharing task states and intermediate results.
  6. 6. The system of claim 1, wherein updating the side weights associated with the corresponding transportation segments in the multi-segment intermodal transportation network map based on the segment risk parameters results in a risk weighted transportation network map, comprising: acquiring basic side weights corresponding to all sides in the multi-mode intermodal transportation network diagram, wherein the basic side weights are determined by basic transportation parameters associated with corresponding transportation road sections; generating weight increment parameters associated with corresponding transportation road segments based on the road segment risk parameters, wherein the weight increment parameters represent the influence degree of external risk information on the basic side weight; according to a preset weighting updating rule, carrying out fusion calculation on the weight increment parameter and the basic side weight to obtain an updated side weight, and writing the updated side weight into the multi-type intermodal transportation network diagram to form a risk weighting transportation network diagram; And when the external risk information changes, carrying out consistency update on the risk weighted transportation network diagram based on the change amount of the updated side weight, so that the risk weighted transportation network diagram is kept corresponding to the current external risk information.
  7. 7. The system of claim 1, wherein said constructing a combined optimization model on said risk weighted transportation network graph that satisfies said transportation constraints and solving, outputting a transportation path plan corresponding to said demand parameters, comprises: Determining decision variables used for representing transportation road section selection and transportation resource allocation based on the risk weighted transportation network diagram, and taking cost parameters, aging parameters and risk parameters associated with each transportation road section as target parameters of the combined optimization model; constructing a constraint set comprising flow conservation constraints and constraint conditions corresponding to the transportation constraint conditions, and writing the constraint set into the combined optimization model to limit the transportation path scheme to meet the transportation constraint conditions; And solving the combined optimization model based on the target parameters and the constraint set to obtain at least one transportation path scheme.
  8. 8. The system of claim 1, wherein the driving the multi-agent collaborative risk processing module to update road segment risk parameters for the affected set of transportation road segments comprises: Acquiring change data of external risk information, and identifying a changed target risk event based on a preset risk triggering rule; Establishing an association relation between an event and a road section based on the time information and the space range information of the target risk event, and determining the affected transportation road section set by combining transportation mode information and exposure parameters associated with cargo sensitivity and carrier redundancy; and generating an updated road segment task set aiming at the affected transportation road segment set, and driving a plurality of agents to execute risk information acquisition and risk parameter generation processing in parallel so as to update road segment risk parameters associated with the affected transportation road segment set.
  9. 9. The system of claim 1, wherein the driving the risk mapping module to update the edge weights and the driving the combined optimization solution module to re-solve to update the transportation path solution comprises: Determining weight increment parameters of sides corresponding to the affected transportation road segment set based on the updated road segment risk parameters, and calling the risk mapping module to execute side weight update on sides corresponding to the affected transportation road segment set in the multi-mode intermodal transportation network diagram so as to obtain an updated risk weighted transportation network diagram; Determining an affected subgraph based on the updated risk weighted transportation network graph, and executing local update on the candidate path set in the affected subgraph to obtain an updated candidate path set; And reconstructing or updating a combined optimization model based on the updated risk weighted transportation network diagram and the updated candidate path set and solving the combined optimization model to obtain an updated transportation path scheme.
  10. 10. A multi-agent based multi-modal intermodal dynamic path planning method based on the system of any one of claims 1 to 9, comprising: Receiving a starting point, an ending point and demand parameters associated with transportation constraint conditions, constructing a multi-mode intermodal transportation network diagram, and generating a candidate path set based on the multi-mode intermodal transportation network diagram; Mapping the transportation road segments associated with the candidate path set into a road segment task set, and driving a plurality of agents to execute risk information acquisition and risk parameter generation processing on the road segment task set in parallel so as to obtain road segment risk parameters associated with each transportation road segment; Updating the side weights associated with the corresponding transportation road sections in the multi-type intermodal transportation network diagram based on the road section risk parameters to obtain a risk weighted transportation network diagram; constructing a combined optimization model meeting the transportation constraint condition on the risk weighted transportation network diagram, solving the combined optimization model, and outputting a transportation path scheme corresponding to the demand parameter; When external risk information changes, determining an affected transportation road section set, updating road section risk parameters corresponding to the affected transportation road section set, updating the side weight based on the updated road section risk parameters, and re-solving the combined optimization model to update the transportation path scheme; And outputting the transportation path scheme and risk information associated with the transportation path scheme.

Description

Multi-agent-based multi-intermodal dynamic path planning system and method Technical Field The application relates to the technical field of intelligent supply chains, in particular to a multi-agent-based multi-intermodal dynamic path planning system and method. Background In the global supply chain system, transnational freight is generally required to undergo multi-mode intermodal transportation such as maritime transportation, railways, highways, inland rivers, and the transportation path is influenced by various uncertain factors such as weather disasters, port congestion, policy adjustment, land-edge conflict, traffic control, carrier performance fluctuation and the like, and the transportation environment presents remarkable dynamics and complexity. The conventional multi-joint transportation path planning technology mainly comprises a static shortest path model for calculating a shortest path of time or cost based on a deterministic graph algorithm, a multi-objective optimization model for comprehensively considering indexes such as cost, aging, carbon row and the like, a prediction optimization method for predicting transportation duration or congestion probability by utilizing historical data and adjusting path selection according to the transportation duration or congestion probability, and an optimization method for processing multi-constraint path search by adopting heuristic algorithms such as a genetic algorithm, an ant colony algorithm and the like. However, the method is generally based on static cost or static prediction, and lacks of real-time sensing and dynamic updating capability on sudden risks, so that a planned path is easy to lose effectiveness due to events such as seal navigation and extreme weather in the execution process, meanwhile, multi-source risk information exists in unstructured forms such as texts, automatic identification, structured quantification and side level mapping are difficult to achieve by the existing system, an interpretable and traceable risk model is difficult to form, in addition, the traditional algorithm has contradiction between global optimality and instantaneity, dynamic re-planning requirements of minute level are difficult to support, and stable organization and decision transparency of cross-border multi-modal tasks are further affected. Disclosure of Invention In view of the above, the embodiment of the application provides a multi-agent-based multi-intermodal dynamic path planning system and method, which are used for solving the problems of difficult structured quantification of multi-source risks, difficult dynamic update of side weights and difficult rapid re-planning of paths in the prior art. The first aspect of the embodiment of the application provides a multi-agent-based multi-modal intermodal dynamic path planning system, which comprises a network construction module, a multi-agent collaborative risk processing module, a risk mapping module, a combined optimization solving module, a dynamic re-planning control module and a risk optimization solving module, wherein the network construction module is used for receiving a starting point, an ending point and demand parameters related to transportation constraint conditions, constructing a multi-modal transportation network diagram, generating a candidate path set based on the multi-agent intermodal transportation network diagram, mapping transportation road sections related to the candidate path set into a road section task set, driving a plurality of agents to execute risk information acquisition and risk parameter generation processing on the road section task set in parallel to obtain road section risk parameters related to each transportation road section, the risk mapping module is used for updating side weights related to corresponding transportation road sections in the multi-modal transportation network diagram based on the road section risk parameters to obtain a risk weighted transportation network diagram, combining and solving the combined optimization model meeting the transportation constraint conditions on the risk weighted transportation network diagram, outputting a transportation path scheme corresponding to the demand parameters, and the dynamic re-planning control module is used for determining the affected road section set when external risk information changes, driving the multi-agent collaborative processing module to execute risk information acquisition and risk parameter generation processing on the road section task set to obtain road section risk parameters related to each transportation road section, the driving the updating scheme and outputting the risk optimization scheme. The second aspect of the embodiment of the application provides a multi-agent-based multi-intermodal dynamic path planning method based on a first aspect system, which comprises the steps of receiving a starting point, a finishing point and demand parameters associated with transportation con