Intelligent decision-making method and system based on swarm agent and trusted human-machine hybrid intelligence

Abstract

The application provides a kind of trust man-machine hybrid intelligent decision method and system based on swarm agent, it is related to information technology field, the method is based on hierarchical cascade architecture executes, including: by first-level decision-making agent generates decision scheme and issues execution plan, the process contains the first quantitative evaluation of contingency adaptability, resource trust controllability degree;Receive the empowerment application initiated by second-level decision-making agent because of judging that cannot independently handle emergency event, and execute the second quantitative evaluation containing the influence of event disposal and plan execution effect;In response to the application and authorized by the commander, dynamically configure temporary command authority and / or supplement resource interface for the applicant, update its role identifier after configuring authority;Recycle the authority and resource configured after event processing and restore its original role;Monitoring evaluation results, when evaluation results are lower than corresponding threshold value, interrupt the process and request the intervention of the commander.The application realizes cross-level cooperation, dynamic security adjustment of decision-making authority and whole-process trust controllability.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a trusted human-machine hybrid intelligent decision-making method and system based on swarm intelligence. Background Technology

[0002] With the deep integration of the Internet of Things with advanced modeling technologies, key areas such as emergency response and command decision-making are increasingly reliant on swarm intelligence collaborative systems.

[0003] Currently, there are numerous studies and applications regarding multi-agent collaboration and task allocation. Specifically, existing multi-agent collaboration schemes suffer from several issues. At the architectural level, the diverse designs of current schemes in terms of cross-level linkage and resource integration lead to low overall system collaboration efficiency and data silos. At the dynamic response level, the dynamic adjustment mechanism for decision-making authority is still imperfect when dealing with emergencies, resulting in vulnerabilities in authority transfer and risks of resource misallocation. At the trust assurance level, there is a lack of a systematic quantitative evaluation framework for the trustworthiness and controllability of human-machine hybrid decision-making processes, making it difficult to guarantee the controllability of the entire decision-making process and the credibility of the results.

[0004] Therefore, there is an urgent need for a swarm intelligence decision-making scheme that can achieve efficient cross-level collaboration, support dynamic and secure adjustment of decision-making authority, and ensure that the entire process is trustworthy and controllable. Summary of the Invention

[0005] This invention provides a trusted human-machine hybrid intelligent decision-making method and system based on swarm intelligence, which addresses the shortcomings of existing swarm intelligence systems, such as low efficiency of cross-level collaboration, rigid adjustment of decision-making authority and lack of security, as well as insufficient credibility of the decision-making process and results.

[0006] This invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. The method first executes based on a hierarchical cascaded architecture comprising primary and secondary decision-making agents. The primary decision-making agent generates a decision scheme and distributes the execution plan derived from the decomposition of the decision scheme to the secondary decision-making agents. The process of generating the decision scheme includes a first quantitative assessment, which at least includes a predictive assessment of the suitability of the plan and a matching assessment of the credibility and controllability of the resources involved in the decision scheme. The method receives an empowerment request initiated by the secondary decision-making agent during the execution of the execution plan after it autonomously determines that the agent cannot handle the emergency independently, and performs a second quantitative assessment, which at least includes an execution effect assessment of the impact of the emergency on the handling of the emergency and its impact on the execution effect of the plan. In response to the empowerment request and with authorization from the commander, the method dynamically configures temporary command and decision-making permissions and / or supplementary resource and capability interfaces for the secondary decision-making agent that submitted the empowerment request, and updates its role identifier after configuring the temporary command and decision-making permissions. After the emergency is handled, the temporary command and decision-making permissions and supplementary resource and capability interfaces dynamically configured for the secondary decision-making agent are revoked, and its role identification is restored to its original state; the evaluation results of the first or second quantitative assessment are monitored, and when the evaluation result is lower than the corresponding threshold, the current process is interrupted or suspended, and the commander is requested to intervene.

[0007] This invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. The first-level decision-making intelligence agents include a planning intelligence agent, a preparation intelligence agent, an execution monitoring intelligence agent, and an evaluation intelligence agent. First-level decision-making intelligence agents at the same level support sequential collaboration in a ring-like cascade, while first-level decision-making intelligence agents at different levels support cross-level vertical linkage collaboration. Second-level decision-making intelligence agents are used to: execute the execution plan issued by the first-level decision-making intelligence agents; or, when detecting a sudden event that cannot be handled independently, initiate an empowerment request to obtain temporary command and decision-making authority and / or supplementary resource and capability interfaces required for parallel collaborative execution with other second-level decision-making intelligence agents at the same level; or, after obtaining temporary command and decision-making authority and updating their role identifier to that of a temporary first-level decision-making intelligence agent, organize other intelligence agents to collaboratively handle the sudden event.

[0008] This invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. The evaluation agent has a first working mode and a second working mode. The first working mode is an evaluation sub-module, which is embedded in the planning agent, the preparation agent, and the execution monitoring agent, respectively. The second working mode is an independent evaluation agent.

[0009] This invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. In a first working state, an evaluation submodule embedded in the planning agent is used to perform a first quantitative evaluation of the adaptability of the contingency plan; an evaluation submodule embedded in the preparation agent is used to perform a first quantitative evaluation of the matching evaluation of the credibility, controllability, and task execution controllability of the resources involved in the decision-making scheme; an evaluation submodule embedded in the execution monitoring agent is used to perform a second quantitative evaluation of the impact of handling emergencies and their impact on the plan execution effect. An independent evaluation agent is used to perform a third quantitative evaluation of the overall completion effect of the decision-making scheme.

[0010] This invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. The planning intelligence agents include a superior planning intelligence agent and subordinate planning intelligence agents. The steps for generating decision schemes include: receiving tasks through the superior planning intelligence agent and issuing pre-commands to the subordinate planning intelligence agents in conjunction with the commander's intentions; generating multiple sets of contingency plans in parallel through the superior and subordinate planning intelligence agents; simulating and deducing each set of contingency plans; and determining the decision scheme based on the commander's instructions.

[0011] This invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. The preparation agents include a superior preparation agent and subordinate preparation agents. The execution plan generation step includes: the superior preparation agent performing resource grouping and evaluation according to the decision scheme, generating a local execution plan, and issuing it; and the subordinate preparation agents receiving the local execution plan, generating an adapted branch plan, and issuing it to the corresponding secondary decision-making agents.

[0012] According to the present invention, a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents is provided. The method includes the following steps for dynamically configuring a secondary decision-making intelligence agent that submits an empowerment request: evaluating the empowerment request by executing a monitoring intelligence agent and generating an authorization proposal; after receiving a confirmation instruction from a commander on the authorization proposal, dynamically configuring temporary command and decision-making permissions and / or supplementary resource and capability interfaces for the secondary decision-making intelligence agent that submitted the empowerment request; and updating its role identifier after configuring the temporary command and decision-making permissions.

[0013] The present invention provides a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents. The method further includes: monitoring and interactively verifying the collaborative state of intelligence agents based on an inter-agent communication protocol with added trusted value and controllable value information fields; and verifying the trustworthiness of external resources based on a resource call protocol with added identity authentication and call auditing mechanisms. The external resources include external models, external databases, and external tools.

[0014] This invention provides a trusted human-machine hybrid intelligent decision-making system based on swarm intelligence. The system includes: a first-level decision-making agent configured to execute the operations performed by the first-level decision-making agent in any of the aforementioned methods; a second-level decision-making agent configured to execute the operations performed by the second-level decision-making agent in any of the aforementioned methods; a trusted and controllable module deployed among the agents, configured to monitor and interactively verify the agent collaboration state based on an inter-agent communication protocol with added trusted and controllable value information fields; and a resource management module deployed between the agents and external resources, configured to verify the trustworthiness of external resources based on a resource access protocol with added identity authentication and access auditing mechanisms.

[0015] This invention provides a trusted human-machine hybrid intelligent decision-making system based on swarm intelligence agents. The first-level decision-making intelligence agents include a planning intelligence agent, a preparation intelligence agent, an execution monitoring intelligence agent, and an evaluation intelligence agent. The evaluation intelligence agent has a first working mode and a second working mode; the first working mode is an evaluation sub-module, embedded in the planning intelligence agent, the preparation intelligence agent, and the execution monitoring intelligence agent respectively; the second working mode is an independent evaluation intelligence agent. First-level decision-making intelligence agents at the same level support sequential collaboration in a ring-like cascade, and first-level decision-making intelligence agents at different levels support cross-level vertical linkage collaboration. Second-level decision-making intelligence agents are used to: execute the execution plan issued by the first-level decision-making intelligence agents; or, when detecting a sudden event that cannot be handled independently, initiate an empowerment request to obtain temporary command and decision-making authority and / or supplementary resource and capability interfaces required for parallel collaborative execution with other second-level decision-making intelligence agents at the same level; or, after obtaining temporary command and decision-making authority and updating their role identifier to a temporary first-level decision-making intelligence agent, organize other intelligence agents to collaboratively handle the sudden event.

[0016] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence as described above.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence as described above.

[0018] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence as described above.

[0019] This invention provides a trusted human-machine hybrid intelligent decision-making method and system based on swarm intelligence agents. Through a combination of hierarchical architecture and quantitative evaluation, it achieves end-to-end enhancement from static planning to dynamic execution. A first quantitative evaluation is performed simultaneously during the solution generation stage, preemptively eliminating solutions with poor adaptability or unreliable resources, ensuring the reliability of the decision-making basis from the outset. When unexpected events occur during execution, a second quantitative evaluation is triggered, quantifying the impact of the event in real time and providing accurate and objective decision-making basis for subsequent dynamic adjustment of permissions. Based on this, the system, supported by the evaluation results, completes temporary permission and interface configuration for the executor with personnel authorization and subsequent revocation, making permission adjustment no longer a rigid preset but a data-driven, secure, and controllable elastic response capability. Finally, continuous monitoring of the two evaluation results and setting intervention thresholds constitute a double layer of trust insurance, ensuring that any decline in trustworthiness at any stage can be promptly captured and handed over to personnel for management. This achieves efficient cross-level collaboration, dynamic and secure adjustment of decision-making permissions, and a trusted and controllable decision-making process. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the structure of a trusted human-machine hybrid intelligent decision-making system based on swarm intelligence provided by the present invention.

[0022] Figure 2 This is a flowchart illustrating a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by the present invention.

[0023] Figure 3 This is a complete flowchart of a trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by the present invention.

[0024] Figure 4 This is a schematic diagram of the physical structure of an electronic device provided by the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] In a typical swarm intelligence decision-making system architecture, multiple agents are usually deployed on computing nodes with hierarchical relationships, and exchange information and collaborate on tasks through specific communication protocols (such as Agent to Agent Protocol (A2A) and Autonomous Network Protocol (ANP)). The commander accesses the system through a dedicated command and control terminal to issue tasks and monitor their execution.

[0027] When dealing with complex tasks, although multiple agents may logically belong to the same task system, they are usually treated as independent modules that operate according to a fixed process in application logic. Their decision-making process may be based on independent calculations of their own models or may require task scheduling and result aggregation through a unified control node.

[0028] In this architecture, task execution and permission allocation typically follow static, predefined rules. When an execution-type intelligent agent encounters unexpected events outside its authority or capabilities during a task, the existing collaboration process cannot support its dynamic acquisition of higher decision-making authority and additional resources. Commanders or higher-level systems must manually assess and operate to re-decompose tasks, grant permissions, and allocate resources, resulting in inefficiency.

[0029] Furthermore, some existing technologies involving multi-agent collaboration focus primarily on task allocation optimization or communication efficiency improvement. They fail to conduct real-time quantitative assessments of the credibility and controllability of the agent's own state, interactive behavior, and external resource calls during execution, and do not deeply couple these assessments with the permission management mechanism. This results in the system lacking proactive and adaptive security intervention capabilities when facing abnormal or untrustworthy situations.

[0030] Meanwhile, if all the complex collaborative decision-making and trust assessment logic is concentrated at the top level or in the cloud, all perceived data and decision instructions will have to go through a long transmission and processing chain, which will make it difficult to meet the stringent requirements of real-time performance, autonomy and reliability in scenarios such as emergency response.

[0031] To address the aforementioned technical problems, this invention provides a trusted human-machine hybrid intelligent decision-making method and system based on swarm intelligence. By constructing a collaborative system of swarm intelligence with clearly defined responsibilities and a ring-shaped cascade within a hierarchical architecture, and based on a quantitative and trustworthy assessment and threshold intervention mechanism that runs through the entire process of task generation, execution, and dynamic adjustment of permissions, cross-level collaboration, dynamic and secure adjustment of decision-making permissions, and full-process trustworthiness and controllability are achieved.

[0032] This invention provides a trusted human-machine hybrid intelligent decision-making system based on swarm intelligence agents. The system includes a first-level decision-making agent, a second-level decision-making agent, a trusted and controllable module, and a resource management module.

[0033] In this embodiment of the invention, a first-level decision agent is configured to perform the operations performed by the first-level decision agent in the following method embodiments; a second-level decision agent is configured to perform the operations performed by the second-level decision agent in the following method embodiments.

[0034] For example, a first-level decision-making agent can generate decision-making schemes and issue execution plans; receive and process empowerment requests from second-level decision-making agents; dynamically configure permissions and interfaces for second-level decision-making agents and update their role identifiers after authorization; revoke permissions and interfaces and restore their roles after the task is completed; monitor quantitative evaluation results and request commander intervention when they fall below a threshold.

[0035] In this embodiment of the invention, a trusted and controllable module is deployed among the various intelligent agents. It is configured to monitor and interactively verify the collaborative state of the intelligent agents based on an inter-agent communication protocol that includes trusted and controllable value information fields.

[0036] For example, the Trusted and Controllable Module adds quantitative fields of Trust and Controllability values ​​to basic protocols such as A2A, ANP, and Artificial Society Physical Information System Protocol (ACP) to monitor the collaborative status and task execution progress of intelligent agents in real time, ensuring that the command team can reliably verify and controllably take over the task results of intelligent agents.

[0037] In this embodiment of the invention, a resource management module is deployed between the intelligent agent and external resources. It is configured to verify the trustworthiness of external resources based on a resource invocation protocol with added identity authentication and invocation auditing mechanisms. These external resources include external models, external databases, and external tools.

[0038] For example, the resource management module extends the trusted verification rules for the resource registration, invocation and execution phases based on the Model Call Protocol (MCP) to record and trace the access qualifications, invocation process and execution results of external resources throughout the entire process.

[0039] Thus, by deeply integrating the trusted and controllable mechanism and the dynamic invocation mechanism into the system architecture, this invention constructs a security barrier from two dimensions: agent collaboration and external resource invocation, providing a trusted basis for the dynamic adjustment of decision-making permissions.

[0040] In this embodiment of the invention, the first-level decision-making agent includes a planning agent, a preparation agent, an execution monitoring agent, and an evaluation agent.

[0041] The evaluation agent has a first working mode and a second working mode.

[0042] For example, the first working mode is an evaluation submodule, which is embedded in the planning agent, the preparation agent and the execution monitoring agent respectively; the second working mode is an independent evaluation agent.

[0043] In this embodiment of the invention, sequential collaboration in a ring-shaped cascade is supported between first-level decision agents at the same level, and cross-level vertical collaboration is supported between first-level decision agents at different levels.

[0044] For example, the sequential collaboration of the ring-shaped cascade is manifested as follows: within the same level (such as a certain command level), the planning agent, the preparation agent, the execution monitoring agent, and the evaluation agent can sequentially transmit and process information according to the task flow, and jointly generate and iteratively optimize the emergency support plan; the cross-level linkage collaboration is manifested as follows: the execution monitoring agents of the upper and lower levels can synchronously track the task status and directly exchange control commands.

[0045] In this embodiment of the invention, the secondary decision-making agent can perform the following three operations in different scenarios.

[0046] 1) The second-level decision-making agent is used to execute the execution plan issued by the first-level decision-making agent.

[0047] For example, a secondary decision-making agent (such as a transportation unit agent) receives and parses the plan to transport goods Y along route X, automatically controls the vehicle to complete actions such as driving, loading, and unloading, and reports its location and status in real time.

[0048] 2) Second-level decision-making agents are used to initiate empowerment requests when they detect emergencies that cannot be handled independently, in order to obtain temporary command and decision-making permissions and / or supplementary resource and capability interfaces required for parallel and collaborative execution with other second-level decision-making agents at the same level.

[0049] For example, when the transportation unit agent detects a road interruption, it automatically initiates an application to request temporary permission to direct other backup transportation units on the same road segment and invokes the route planning tool.

[0050] 3) Secondary decision-making agent, which is used to organize other agents to coordinate the handling of emergencies after obtaining temporary command and decision-making authority and updating the role identifier to a temporary primary decision-making agent.

[0051] For example, after obtaining authorization, the agent's role identifier is updated, and it can assign detour tasks to other transportation unit agents in the vicinity and coordinate loading equipment to jointly solve passage obstacles.

[0052] Thus, by defining the flexible role transition of a secondary decision-making agent from execution to temporary command, this invention achieves a flattened, self-organizing, and collaborative response to emergencies.

[0053] The following is combined Figure 1 The above-mentioned trustworthy human-machine hybrid intelligent decision-making system based on swarm intelligence is described in detail.

[0054] like Figure 1 As shown, the trusted human-machine hybrid intelligent decision-making system based on swarm intelligence agents includes a first-level decision-making agent 110, multiple first-level decision-making agents 120, and multiple second-level decision-making agents 130.

[0055] Among them, the superior first-level decision-making agent 110 is connected to multiple subordinate first-level decision-making agents 120, and each subordinate first-level decision-making agent 120 is connected to multiple second-level decision-making agents 130.

[0056] In this embodiment of the invention, the superior first-level decision-making intelligent agent 110 includes a superior planning intelligent agent 111, a superior preparation intelligent agent 112, a superior execution monitoring intelligent agent 113, and a superior evaluation intelligent agent 114.

[0057] For example, the superior planning agent 111 is used to receive the task documents issued by the commander, confirm with the commander the accuracy of the understanding of the intent and the accuracy of the judgment of the corresponding scenario (such as emergency support), issue pre-orders to the subordinate planning agents, generate multiple sets of contingency plans through rule matching, and verify the feasibility of the contingency plans through simulation and deduction.

[0058] Specifically, when generating a contingency plan, the superior planning agent 111 can call the embedded evaluation submodule (i.e., the first working state of the evaluation agent) to predict and evaluate the adaptability of different contingency plans, assist in generating multiple alternative contingency plans that meet the requirements of the task objectives and time windows, and finally submit the verified contingency plan to the commander for selection. Based on the selected contingency plan, a decision-making framework with clear task objectives, task time windows and core execution requirements is generated.

[0059] Thus, this invention ensures the relevance and feasibility of the decision-making scheme by combining the intent confirmation, contingency plan generation, and deduction verification process of the superior planning agent with the adaptability assessment of the embedded evaluation sub-module.

[0060] In this embodiment of the invention, the superior preparation agent 112 is used to group resources such as transportation units, support personnel and route planning models participating in the mission according to the plan selected by the commander, and establish a clear command and organization relationship.

[0061] For example, the superior preparation agent 112 can call the embedded evaluation submodule to evaluate the credibility, controllability, and matching degree of resources such as transportation units and support personnel with the current task. Based on the evaluation results, it can select suitable resources, assist in generating a detailed plan at this level, carry out pre-action preparations (such as resource scheduling, vehicle inspection, etc.), and simultaneously distribute the plan at this level to the subordinate preparation agents.

[0062] Thus, by preparing a resource grouping and optimization mechanism for intelligent agents, combined with resource reliability assessment, this invention provides resource assurance for the implementation of decision-making schemes.

[0063] In this embodiment of the invention, the superior execution monitoring agent 113 is used to monitor the operational status of each action group and the guarantee resources in real time according to the plan generated by the superior preparation agent.

[0064] For example, the superior execution monitoring agent 113 can continuously determine whether there are abnormal situations such as resource conflicts or self-interference and mutual interference, and actively identify external emergencies (such as sudden road control or sudden weather changes in emergency support), conduct assessments based on action monitoring data and external event information, and provide handling suggestions.

[0065] Furthermore, for empowerment requests submitted by subordinates, the superior execution monitoring agent 113 can assess the impact of emergencies on the expected action plan and the availability of resources, determine whether to respond to the request, and if so, simultaneously resolve resource conflicts related to the empowerment request, and complete the authorization empowerment after authorization by the command team.

[0066] Furthermore, the superior-level monitoring agent 113 can also determine whether the task completion status reported by the subordinates meets the requirements of the action plan and provide timely suggestions for task correction.

[0067] Thus, by implementing a full-process monitoring and dynamic response mechanism for the monitoring agent, this invention achieves closed-loop control over the task execution process, providing support for the rapid handling of emergencies.

[0068] In this embodiment of the invention, the superior evaluation agent 114 is responsible for evaluating both the effectiveness of event / task execution and the effectiveness of solution completion.

[0069] In one example, when the first working state (i.e., evaluation submodule) of the superior evaluation agent 114 exists, when it is embedded in the superior planning agent 111, it is evaluation submodule 111a, which assists in completing the prediction evaluation of the adaptability of the contingency plan; when it is embedded in the superior preparation agent 112, it is evaluation submodule 112a, which assists in completing the evaluation of resource credibility and matching degree; when it is embedded in the superior execution monitoring agent 113, it is evaluation submodule 113a, which assists in completing the evaluation of the execution effect of the event / task, and outputs the plan adjustment suggestions based on the evaluation results of the execution effect of the event / task.

[0070] In another example, when the second working mode (i.e., independent agent) of the superior evaluation agent 114 exists, it collects and summarizes the data of the entire action process to evaluate whether the completion effect of the plan meets the preset requirements. If it does, it generates a task summary report; if it does not, it outputs a plan adjustment suggestion and feeds it back to the superior planning agent 111 or the superior execution monitoring agent 113.

[0071] Thus, by employing a dual-form design for evaluating intelligent agents, this invention achieves full-cycle coverage of process evaluation and outcome evaluation, providing data support for decision optimization and process adjustment.

[0072] In this embodiment of the invention, the lower-level first-level decision-making intelligent agent 120 includes a lower-level planning intelligent agent 121, a lower-level preparation intelligent agent 122, a lower-level execution monitoring intelligent agent 123, and a lower-level evaluation intelligent agent 124.

[0073] For example, the lower-level planning agent 121 is used to take the pre-issued command from the higher-level planning agent 111 as task input. It can combine the actual situation of its jurisdiction (such as regional road network conditions, resource warehouse locations, etc.) to understand the higher-level intention, carry out planning work in parallel, formulate the corresponding contingency plan and branch plan, report to the higher-level planning agent 111, receive the conflict detection results and correction prompts from the higher-level agent, and adjust and optimize the contingency plan and branch plan of its own level.

[0074] Thus, this invention achieves hierarchical decomposition and localization of decision-making schemes through the linkage and collaboration of planning agents at different levels, thereby improving the efficiency of cross-level collaboration.

[0075] For example, the lower-level preparation agent 122 receives the local plan issued by the upper-level preparation agent 112, uses it as task input, regroups the resources under its jurisdiction such as transportation units, support personnel, and models, establishes an organizational relationship to adapt to the branch plan, calls the embedded evaluation submodule to evaluate the matching degree between the local resources and the branch plan, selects the appropriate resources to assist in generating the local branch plan, and carries out pre-action preparation, while issuing the branch plan to the corresponding secondary decision-making agent 130.

[0076] Thus, by adapting local resources and decomposing plans for lower-level prepared intelligent agents, this invention ensures the operability of decision-making schemes at the execution level.

[0077] For example, the lower-level execution monitoring agent 123 is used to monitor the execution status of the branch plan at this level and to summarize and analyze the execution data reported by the secondary decision-making agent 130.

[0078] Specifically, the lower-level execution monitoring agent 123 can monitor the real-time status of its own action group and support resources, identify resource conflicts, self-interference and mutual interference, and external emergencies, and provide handling suggestions. For empowerment applications initiated by the secondary decision-making agent 130, the lower-level execution monitoring agent 123 can first conduct a preliminary review, assess the rationality and urgency of the application, and then report it to the superior execution monitoring agent 113. At the same time, based on the task completion status of the secondary decision-making agent 130, it can determine whether it meets the requirements of the branch plan, provide timely corrective suggestions, and if there are problems that exceed the handling capacity of the lower level, it can be reported to the superior execution monitoring agent 113 simultaneously.

[0079] Thus, this invention enables the direct monitoring and management of subordinate secondary decision-making agents by lower-level execution monitoring agents, achieving localized, hierarchical feedback and rapid response to task execution status, and reducing delays caused by communication across multiple levels.

[0080] In this embodiment of the invention, the working mode of the lower-level evaluation agent 124 is consistent with that of the upper-level evaluation agent 114, including the embedded sub-module mode and the independent agent mode.

[0081] In one example, when used as an embedded submodule, it is embedded in the lower-level planning agent 121 as an evaluation submodule 121a; embedded in the lower-level preparation agent 122 as an evaluation submodule 122a; and embedded in the lower-level execution monitoring agent 123 as an evaluation submodule 123a, thereby evaluating the execution effect of each stage at this level in real time.

[0082] In another example, when acting as an independent agent, the lower-level evaluation agent 124 collects and summarizes the execution process data of its own branch plan, evaluates the completion effect of the branch plan, generates its own evaluation report, and reports it to the upper-level evaluation agent 114, providing basic data for global solution evaluation.

[0083] Thus, this invention achieves hierarchical aggregation and global analysis of evaluation data through collaborative evaluation by upper and lower level evaluation agents, ensuring the comprehensiveness of the evaluation results.

[0084] In this embodiment of the invention, the secondary decision-making agent 130 includes multiple agents (such as agents 1 to N).

[0085] Each intelligent agent includes a perception submodule 131, a judgment submodule 132, a decision submodule 133, and an execution submodule 134.

[0086] For example, each agent is used to receive the branch plan issued by the lower-level first-level decision agent 120, parse the plan and automatically generate instruction code, execute the instructions and evaluate the execution effect.

[0087] Specifically, during the execution of instructions, the secondary decision-making agent 130 can identify sudden events through its own perception submodule 131 and determine whether it can handle the situation independently through the judgment submodule 132. If it cannot handle the situation independently, it can initiate an empowerment request to the superior execution monitoring agent (lower execution monitoring agent 123 or superior execution monitoring agent 113) through the decision-making submodule 133. After obtaining authorization and / or empowerment, the secondary decision-making agent 130 may be dynamically configured with one or more of the following supports: temporary command and decision-making authority additions or supplementary resource and capability interfaces.

[0088] For example, if temporary command and decision-making authority is granted, the role identifier of the secondary decision-making agent 130 will be updated to a temporary primary decision-making agent, which can organize other agents to coordinate the handling of tasks; after the task is completed, the authority will be released and the agent will be restored to the role of a secondary decision-making agent.

[0089] For example, if only supplementary resource and capability interfaces are obtained, the secondary decision agent 130 retains its secondary decision agent role and independently completes the task using the newly added resources; after the task is completed, all resource interfaces dynamically configured for the agent are reclaimed.

[0090] Thus, by dynamically switching roles and flexibly adjusting permissions of the two-level decision-making intelligent agents, this invention enables rapid response and efficient handling of emergencies, thereby enhancing the system's dynamic adaptability.

[0091] The following is combined Figure 2 This invention describes a trustworthy human-machine hybrid intelligent decision-making method based on swarm intelligence agents, which is based on... Figure 1 The diagram shows a hierarchical, cascaded architecture that includes first-level and second-level decision-making agents.

[0092] Figure 2 This is a flowchart illustrating the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by the present invention, as shown below. Figure 2 As shown, the method includes the following:

[0093] S201. The first-level decision-making agent generates a decision scheme and distributes the execution plan obtained from the decomposition of the decision scheme to the second-level decision-making agent.

[0094] The first-level decision-making agents include planning agents, preparation agents, execution and monitoring agents, and evaluation agents. First-level decision-making agents at the same level support sequential collaboration in a ring-like cascade, while first-level decision-making agents at different levels support cross-level, vertically linked collaboration.

[0095] In this embodiment of the invention, the evaluation agent has a first working mode and a second working mode.

[0096] For example, the first working mode is an evaluation submodule, which is embedded in the planning agent, the preparation agent and the execution monitoring agent respectively; the second working mode is an independent evaluation agent.

[0097] It should be noted that the descriptions of the aforementioned first-level and second-level decision-making agents, as well as their respective components, can be found above. Figure 1 The corresponding detailed description is omitted here to avoid repetition.

[0098] The secondary decision-making agent is used to: execute the execution plan issued by the primary decision-making agent; or, when an emergency that cannot be handled independently is detected, initiate an empowerment request; or, after obtaining temporary command and decision-making authority and updating its role identifier to a temporary primary decision-making agent, organize other agents to collaboratively handle the emergency.

[0099] Optionally, such as Figure 1 As shown, a planning agent can include a superior planning agent and a subordinate planning agent.

[0100] In some embodiments, a superior planning agent receives a task and issues a pre-command to a subordinate planning agent in accordance with the commander's intent.

[0101] For example, the tasks that the superior planning agency receives can be complex scenarios such as emergency response and joint operations. Taking an emergency support task in a certain area as an example, the task document that the superior planning agency receives clearly states the core objective of delivering emergency supplies and professional forces to multiple dispersed key locations.

[0102] Specifically, the superior planning agent first confirms with the commander that the intention to prioritize key locations, optimize the overall delivery route, and ensure efficient allocation of materials is accurately understood. At the same time, it determines that the current situation of limited road capacity and changeable weather conditions in the mission area is true. Then, it issues a pre-order to the subordinate planning agents (such as the command agents corresponding to each support direction) to activate the joint emergency support plan and carry out multi-path coordinated delivery, so as to ensure that the superior and subordinate have a consistent understanding of the core requirements of the mission.

[0103] Furthermore, multiple contingency plans are generated in parallel by the superior and subordinate planning agents, each plan is simulated and analyzed, and a decision-making scheme is determined based on the commander's instructions.

[0104] For example, the superior planning agent generates multiple plans based on emergency support tasks through rule matching, such as direct air delivery + ground vehicle transportation, establishing forward transit hubs + last-mile distributed delivery, centralized resource scheduling + on-demand dynamic allocation; the subordinate planning agent generates branch plans such as special plans for material delivery along mountain routes and multi-point delivery plans in urban areas, based on the characteristics of its own support direction (such as terrain and distance).

[0105] Specifically, both upper and lower level planning agents verify contingency plans through simulations and exercises. For example, the upper-level planning agent simulates the coverage and timeliness of different delivery methods, while the lower-level planning agent simulates the feasibility and safety risks of transportation routes under specific terrain conditions. During the simulation, the execution difficulties, resource requirements, and expected effects of the contingency plans are recorded simultaneously. Finally, all contingency plans are submitted to the commander, who then selects the target contingency plan that combines air and ground operations, establishes transit stations, and dynamically adjusts the allocation scheme based on the real-time situation in the mission area.

[0106] Thus, this invention efficiently generates and optimizes multiple decision-making plans through parallel operations and simulations of planning agents at different levels, providing commanders with scientific decision-making options.

[0107] Optionally, such as Figure 1 As shown, a preparing agent can include a superior preparing agent and a subordinate preparing agent.

[0108] In some embodiments, a higher-level preparation agent can group and evaluate resources according to the decision-making scheme, generate a local execution plan, and issue it.

[0109] For example, the superior preparation agent, based on the selected target plan, groups the transportation units, loading and unloading teams, support materials and route planning models involved in the task, and establishes a command and organization relationship of "air delivery group - ground transportation group - material allocation group".

[0110] Specifically, the superior-level intelligent agent invokes the embedded evaluation submodule to assess the carrying capacity (reliability), response scheduling degree (controllability), and matching degree with the planned route of each transportation unit. For example, it assesses whether the off-road performance of the heavy transport convoy meets the requirements of mountainous road conditions, whether the range and payload of the drones are suitable for the air delivery scenario, and selects the best resources based on the evaluation results to generate the execution plan of "the air group is responsible for the delivery of materials to points A and B, the ground group travels along routes C and D, and the dispatch group sorts and redistributes materials at the transfer station H".

[0111] At the same time, advance action preparations (such as vehicle maintenance, material loading, and personnel standby) are carried out, and the execution plan is distributed to lower-level prepared agents.

[0112] Furthermore, the lower-level preparation agent can receive the execution plan of this level, generate an adapted branch plan, and distribute it to the corresponding secondary decision-making agent.

[0113] For example, after receiving the plan from the superior, the lower-level preparation agent regroups the resources under its jurisdiction to generate branch plans such as "Northern Route Ground Transportation Plan" and "Southern Route Air Support Plan", and then distributes them to the corresponding secondary decision-making agents.

[0114] Specifically, the lower-level intelligent agents, based on the road conditions and convoy composition of the northern route, refine the "northern route ground transportation plan" into "convoy No. 1 is responsible for the rapid passage of the first section, and convoy No. 2 is responsible for the heavy-load transportation of the second section," and clarify their respective departure times, routes, communication links, and rendezvous points, forming detailed instructions that can be executed immediately.

[0115] Thus, by preparing resource grouping and planning for intelligent agents, and combining resource credibility and matching degree assessment, the present invention transforms the target plan into an executable action plan, ensuring the rationality of resource allocation and the operability of the plan.

[0116] In some embodiments, the process of generating a decision-making scheme includes a first quantitative evaluation.

[0117] For example, the first quantitative assessment includes at least a predictive assessment of the suitability of the contingency plan and a matching assessment of the credibility and controllability of the resources involved in the decision-making plan.

[0118] The evaluation process is based on a pre-set quantitative indicator system (such as suitability score, credibility score, and matching score), with each score ranging from 0 to 100.

[0119] Specifically, the prediction and evaluation of the suitability of the contingency plan is performed by the evaluation submodule embedded in the planning agent. It provides a quantitative score by analyzing indicators such as the degree of fit between the plan and the task objectives and the degree of compatibility with the scenario constraints. For example, a contingency plan has a suitability score of 85 points, which meets the threshold requirement of "suitability score ≥ 80 points".

[0120] Specifically, the matching assessment of resource reliability and controllability is performed by the evaluation submodule embedded in the preparation agent. It provides reliability and controllability scores by analyzing indicators such as the resource's historical execution records (e.g., the on-time rate of past tasks of transportation units) and current status (e.g., vehicle availability). It also provides a task execution controllability score by analyzing the degree of fit between resource functions and task requirements. For example, a heavy transport fleet might have a reliability score of 90, a controllability score of 88, and a task execution controllability score of 92, all meeting the threshold requirements for their respective indicators.

[0121] It should be noted that the inter-agent state monitoring and interaction verification used in the first quantitative evaluation are based on inter-agent communication protocols (A2A, ANP, ACP) with added fields of trusted value and controllable value information. The protocol messages carry the quantitative evaluation results through the added fields to ensure the reliable transmission of evaluation data.

[0122] Thus, this invention uses a first quantitative assessment to conduct a dual screening of decision-making schemes and resources, avoiding the entry of schemes and resources with insufficient adaptability or low credibility into the execution process from the source, thereby improving the reliability of decision-making schemes.

[0123] S202. Receive the empowerment request initiated by the secondary decision-making agent during the execution of the execution plan after it has determined that the emergency is one that cannot be handled independently, and perform the second quantitative assessment.

[0124] For example, during the execution of the branch plan, the secondary decision-making agent monitors the surrounding environment and task execution status in real time through its own perception submodule. When it detects an emergency that exceeds its own handling capacity, it automatically triggers the empowerment application process.

[0125] Specifically, taking the secondary decision-making agent that executes the northern route ground transportation plan as an example, it detects an emergency event during transportation where "the main road ahead is impassable due to an unforeseen situation, and the original route is blocked." By analyzing its own capabilities through the judgment submodule, it finds that it only has the execution authority to transport according to the planned route, but no route replanning decision-making authority or the right to dispatch other transportation forces in the relevant area, and cannot solve the problem independently. Subsequently, it conducts a quantitative assessment of the impact of the emergency event handling through the evaluation submodule (second quantitative assessment).

[0126] For example, assess the impact of "the delay in the delivery of materials due to route interruption on the overall progress of the guarantee" and give an impact score of 75 points (the value range is 0-100 points, the higher the score, the greater the impact). If the condition of "impact score ≥ 70 points, an empowerment application should be initiated" is met, and then an empowerment application should be initiated to the superior execution monitoring intelligent agent.

[0127] The application may include a description of the emergency, an assessment of its own handling capabilities, the required decision-making authority (such as the right to plan alternative routes and the right to schedule collaborative units), and resource interface requirements.

[0128] Thus, this invention ensures the relevance and necessity of empowerment applications by autonomously identifying and assessing the capabilities of a secondary decision-making agent for sudden events, combined with the impact assessment of a second quantitative evaluation.

[0129] In this embodiment of the invention, the second quantitative assessment includes at least an assessment of the impact of the handling of the emergency and its impact on the effectiveness of the plan.

[0130] For example, the second quantitative assessment is performed by the assessment submodule embedded in the execution monitoring agent. The assessment indicators include the impact of the unexpected event on the original planned execution progress, the impact on the achievement of the task objectives, the impact on the surrounding environment and other actions, etc., and a comprehensive impact score is obtained by weighted calculation.

[0131] Specifically, after the superior execution monitoring intelligent agent receives the empowerment application, it calls the embedded evaluation submodule to further refine the evaluation based on the application content and the original plan requirements: the road interruption will cause a transportation delay of 2 hours, affecting the receipt of materials at two key points (progress impact score 80 points); the delay may cause the subsequent operation at the point to be suspended, affecting the overall task rhythm (target impact score 78 points); there are no other equivalent routes that can be replaced immediately, and it is necessary to coordinate with other nearby transportation units to detour and provide support, which may affect their original tasks (related impact score 72 points). After weighted calculation, the comprehensive impact score is 77 points, which meets the threshold requirement of "comprehensive impact score ≥ 75 points requires response to empowerment application".

[0132] It should be noted that in the second quantitative evaluation process, the interaction verification between intelligent agents is also based on the A2A, ANP, and ACP protocols with added trusted value and controllable value fields to ensure the authenticity of the evaluation data and the security of transmission.

[0133] Thus, this invention objectively assesses the impact of emergencies through a second quantitative evaluation, providing a quantitative basis for deciding whether to respond to empowerment applications and avoiding the blind adjustment of permissions.

[0134] S203. In response to the empowerment request and with the authorization of the commander, dynamically configure temporary command and decision-making permissions and / or supplementary resource and capability interfaces for the secondary decision-making agent that submitted the empowerment request, and update its role identifier after configuring the temporary command and decision-making permissions.

[0135] In some embodiments, an empowerment request can be evaluated by a monitoring agent to generate an authorization proposal.

[0136] For example, the execution monitoring agent (upper or lower level) reviews the rationality of the empowerment application, the necessity of the required permissions, and the availability of the resource interface based on the second quantitative evaluation results, the current resource configuration status, and the task priority.

[0137] Specifically, the application for empowerment of the monitoring agent analysis requires the right to plan alternative routes and the right to schedule collaborative units. It is confirmed that this right belongs to the command-type right of the first-level decision-making agent, and that there are other schedulable transportation unit resources and the interface can be called normally. At the same time, combined with the task priority (emergency support is the high priority), it is determined that granting this right will not affect the execution of other core tasks. Thus, an authorization proposal is generated: "Agree to empowerment, grant the right to plan alternative routes and the right to schedule collaborative units, open the relevant transportation unit resource interface, and the empowerment is valid until the alternative route is opened and the materials are delivered to the target location."

[0138] Furthermore, upon receiving confirmation instructions from the commander regarding the authorized proposal, the secondary decision-making agent that submitted the empowerment request is dynamically configured with temporary command and decision-making permissions and / or supplementary resource and capability interfaces.

[0139] For example, the execution monitoring agent submits the authorization proposal to the commander, who then views the application details, evaluation results, and authorization proposal through the command terminal. After confirming that everything is correct, the commander issues the "authorize" instruction.

[0140] Specifically, after the monitoring agent receives the commander's instructions, it dynamically configures the command and decision-making permissions for the secondary decision-making agent and updates its role identifier to "temporary primary decision-making agent", enabling it to have the core functions of a command-type agent; at the same time, it opens the corresponding resource interface, enabling it to call external resources such as geographic information systems (for planning alternative routes) and surrounding transportation unit scheduling systems.

[0141] It should be noted that the process of calling external resources is based on a resource calling protocol (such as MCP) with added identity authentication and calling auditing mechanisms. When calling resources, the secondary decision-making agent needs to be authenticated (based on a preset trusted identifier). The system automatically records audit information such as calling time, calling content, and execution results to ensure the traceability of resource calls.

[0142] For example, the empowered secondary decision-making agent queries the surrounding road network through the geographic information system interface and plans an alternative route to "detour via backup road X". It coordinates a nearby reserve convoy through the transportation unit scheduling interface and issues a scheduling instruction to "coordinate the transportation task of the detour section". At the same time, it adjusts the travel sequence of itself and other related units to ensure that the materials are delivered as soon as possible.

[0143] Furthermore, after configuring temporary command and decision-making permissions, update their role identifiers.

[0144] For example, the role identifier is updated synchronously in the system permission management center. The status of the secondary decision-making agent in the system changes from "executor" to "temporary command agent". The collaborative network it accesses, the types of instructions it can receive, and the permissions to issue instructions are all adjusted accordingly.

[0145] Specifically, after the system updates its identifier, other secondary decision-making agents at the same level can recognize its new role and receive its instructions for organization and coordination; higher-level agents can also see the change in its role status in the monitoring interface, which facilitates overall control.

[0146] Thus, this invention achieves precise and secure configuration of permissions and resources through a closed-loop process of review-authorization-configuration, combined with the trusted invocation mechanism of the MCP protocol, enabling the secondary decision-making agent to quickly acquire the ability to handle emergencies.

[0147] S204. After the emergency is handled, reclaim the temporary command and decision-making authority and supplementary resource and capability interfaces that were dynamically configured for the secondary decision-making agent, and restore its role identifier to its original state.

[0148] In some embodiments, after the secondary decision-making agent (temporary primary decision-making agent) completes the handling of an emergency, it automatically reports the task completion status to the execution monitoring agent that it has authorized to empower.

[0149] The task completion status may include the results of handling emergencies, permission usage, resource call records, and a task summary.

[0150] For example, after completing the tasks of "alternative route planning", "cooperative unit scheduling" and "material transportation", the empowered secondary decision-making agent reports a summary that clearly states "the alternative route has been opened, the materials have been delivered to the target location, the overall progress delay is controlled within 1 hour, and the permissions and resources have not been used beyond the scope".

[0151] Specifically, after receiving the reported information, the monitoring agent calls the embedded evaluation submodule to assess the handling effect, confirms that the emergency has been properly resolved and the task completion effect meets the requirements, and then sends a "permission revocation" command to automatically revoke the command and decision-making permissions configured for the agent, close the corresponding resource interface, restore its role identifier to a level two decision-making agent, and update the permission change log and resource call audit record to form a complete closed loop.

[0152] Thus, by automatically revoking permissions and restoring roles after a task is completed, this invention ensures the temporary nature and controllability of permission adjustments, prevents security risks caused by long-term permission retention, and achieves traceability of permission changes through log recording.

[0153] S205. Monitor the evaluation results of the first or second quantitative assessment. When the evaluation result is lower than the corresponding threshold, interrupt or suspend the current process and request the commander to intervene.

[0154] In this embodiment of the invention, the threshold for the first quantitative evaluation may include a scheme suitability threshold (e.g., 80 points), a resource credibility threshold (e.g., 85 points), and a resource matching threshold (e.g., 80 points); the threshold for the second quantitative evaluation includes a comprehensive impact threshold (e.g., 75 points).

[0155] The threshold can be preset or dynamically adjusted by the commander according to the task type and scenario requirements.

[0156] For example, during the process of generating a decision-making plan, if the adaptability quantitative evaluation score of a certain plan is 75 points, which is lower than the preset threshold of 80 points, the decision-making plan generation process will be automatically interrupted. The evaluation report, score details and improvement suggestions of the plan will be submitted to the commander, requesting the commander to intervene and determine whether to continue to optimize the plan or regenerate the plan.

[0157] Specifically, after reviewing the assessment report, the commander found that the plan's lack of adaptability stemmed primarily from "not fully considering the constraints of nighttime traffic capacity in the mission area." Subsequently, an order was issued to "supplement the nighttime action plan and re-evaluate the scenario." The superior planning agency combined the order with optimization of the plan, and conducted a new simulation and quantitative assessment until the score met the threshold requirements.

[0158] In another exemplary scenario, the comprehensive impact score of the second quantitative assessment is 72 points, which is lower than the preset threshold of 75 points. The empowerment application response process is suspended, and the assessment results and application details are submitted to the commander. The commander judges that the impact of the emergency is limited and can be resolved by the secondary decision-making agent through existing resources without granting command-type permissions. Then, the commander issues an instruction to "reject the empowerment application and guide the secondary decision-making agent to try to smooth things over using local resources or wait briefly" and feeds the instruction back to the agent.

[0159] It should be noted that when critical operations such as adjusting decision-making authority or allocating high-risk resources are involved, even if the quantitative assessment results meet the threshold requirements, an intervention reminder can be sent to the commander. The commander will then make the final confirmation before proceeding with the subsequent procedures, ensuring the safety of human-machine collaboration.

[0160] Thus, this invention constructs a dual control mode of automated process + human intervention through quantitative evaluation and threshold intervention mechanism, which not only ensures process efficiency in normal scenarios, but also introduces human judgment at key nodes or in abnormal situations, ensuring the controllability of the entire decision-making process.

[0161] In the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by this invention, the combination of a hierarchical architecture and quantitative evaluation achieves end-to-end enhancement from static planning to dynamic execution. A first quantitative evaluation is performed simultaneously during the solution generation stage, which can preemptively eliminate solutions with poor adaptability or unreliable resources, ensuring the reliability of the decision-making basis from the outset. When unexpected events occur during execution, a second quantitative evaluation is triggered, which can quantify the impact of the event in real time, providing accurate and objective decision-making basis for subsequent dynamic adjustment of permissions. Based on this, the system uses the evaluation results as support, and with personnel authorization, completes the temporary permission and interface configuration of the executor and subsequent revocation, making permission adjustment no longer a rigid preset, but a data-driven, secure, and controllable elastic response capability. Finally, continuous monitoring of the two evaluation results and setting intervention thresholds constitute double trust insurance, ensuring that any decline in trustworthiness at any stage can be captured in a timely manner and handed over to personnel for takeover, thereby achieving efficient cross-level collaboration, dynamic and secure adjustment of decision-making permissions, and trusted and controllable decision-making throughout the entire process.

[0162] Optionally, as described in S201 above, the evaluation agent has a first working mode and a second working mode.

[0163] In one alternative implementation, in the first working state, an evaluation submodule embedded in the planning agent is used to perform a predictive evaluation of the suitability of the plan in the first quantitative evaluation.

[0164] For example, the evaluation submodule embedded in the planning agent presets multi-dimensional adaptability evaluation indicators, including task goal fit, scenario constraint compatibility, resource requirement feasibility, and execution difficulty controllability.

[0165] Specifically, in the evaluation of emergency support plans, the task objective fit index assesses the coverage of the plan to the core objectives of "timely and sufficient delivery of materials and efficient allocation of resources"; the scenario constraint compatibility index assesses the adaptability of the plan to the task area constraints such as "road conditions and weather changes"; the resource demand feasibility index assesses the availability of the resources required by the plan; and the execution difficulty controllability index assesses the controllability of risks during the execution of the plan. Each index is weighted according to preset weights to obtain a quantitative score for adaptability.

[0166] In another alternative implementation, in the first working state, an evaluation submodule embedded in the preparing agent is used to perform a matching evaluation of the credibility, controllability, and task execution controllability of the resources involved in the decision scheme in the first quantitative evaluation.

[0167] For example, the evaluation submodule embedded in the preparation agent sets differentiated evaluation indicators for different types of resources such as transportation units, support personnel, and route planning models: the reliability indicators for transportation units include past task on-time rate and equipment integrity rate, and the controllability indicators include remote scheduling success rate and fault response speed; the reliability indicators for support personnel include professional skill certification level and historical task evaluation, and the controllability indicators include command response speed and cooperation level; the reliability indicators for model resources include route prediction accuracy and historical error rate, and the controllability indicators include parameter adjustment flexibility and anomaly handling capability.

[0168] Specifically, when evaluating a heavy transport fleet (transportation unit resource), its past long-distance transport mission on-time rate was 95% (credibility index score of 95 points), its response time to accepting long-distance route replanning was 30 minutes (controllability index score of 90 points), and its off-road transport function was perfectly matched with the current mission requirements (matching degree score of 100 points). Based on this, the overall evaluation result of the resource meets the threshold requirements.

[0169] In another alternative implementation, in the first working state, an evaluation submodule embedded in the execution monitoring agent is used to perform the second quantitative evaluation of the impact of the handling of the emergency and its impact on the execution effect of the plan.

[0170] For example, the evaluation indicators of this evaluation submodule include progress impact, target impact, related impact, risk diffusion impact, etc. Each indicator is assigned a weight according to the priority of the task scenario. For example, in the emergency support scenario, the target impact (the impact on the overall task rhythm and the effectiveness of site support) has the highest weight.

[0171] Specifically, when assessing a "road disruption" emergency, the progress impact indicator assesses the delay time of the transportation plan, the target impact indicator assesses the degree of impact on the material receiving target at key locations, the correlation impact indicator assesses the linkage impact on other parallel transportation tasks, and the risk diffusion impact indicator assesses whether it may cause traffic congestion or secondary accidents. The comprehensive impact score is obtained through weighted calculation.

[0172] In another alternative implementation, in the second working state, an independent evaluation agent is used to perform a third quantitative evaluation of the overall effectiveness of the decision-making scheme.

[0173] For example, after the task is fully completed, an independent evaluation agent is launched, which summarizes all data generated throughout the entire process from planning, preparation, execution to monitoring, including the raw scores of quantitative evaluation at each stage, commander's instructions, actual resource usage logs, and the final status of task achievement.

[0174] Specifically, the third quantitative assessment takes a holistic perspective and sets up multi-dimensional performance evaluation indicators, such as "task target achievement rate," "plan execution efficiency ratio," "resource utilization efficiency," and "system collaboration credibility." For example, it calculates the target achievement rate by comparing "actual delivery volume / planned delivery volume" and analyzes execution efficiency by comparing "actual time consumption / contingency plan simulation time," thereby generating an objective and comprehensive task summary report for iterative optimization of future decision-making plans and evaluation models.

[0175] Thus, this invention achieves closed-loop verification and feedback of task effectiveness through a third quantitative evaluation, continuously improving the system's decision-making and execution capabilities.

[0176] Optionally, in addition to performing the first quantitative evaluation and / or the second quantitative evaluation as described in S201 and S202 above, the agent collaboration status can also be monitored and interactively verified based on an agent-to-agent communication protocol that adds fields for trusted and controllable values.

[0177] In some embodiments, the message structure of the inter-agent communication protocol (A2A, ANP, ACP) adds trusted value and controllable value fields. The field length is 1 byte and the value range is 0-100 (corresponding to percentages), which carry the quantitative evaluation results of the agent in real time.

[0178] For example, when a higher-level planning agent sends a pre-command to a lower-level planning agent, the A2A protocol message adds a field carrying information such as "higher-level planning agent trust value: 98 points, controllable value: 99 points" and "pre-plan adaptability trust value: 85 points". After receiving the message, the lower-level planning agent judges the trustworthiness of the information by verifying the field values. If the trust value is lower than 80 points, it sends a message to the higher-level agent saying "information verification error, request resend".

[0179] Specifically, the protocol also supports encrypted transmission of field values, using symmetric encryption algorithms to encrypt trusted and controllable value fields to prevent data tampering and ensure the security and reliability of interactions between intelligent agents.

[0180] Thus, by extending the inter-agent communication protocol, this invention integrates quantitative evaluation results into the communication process, enabling real-time transmission and secure verification of evaluation data, and providing technical support for the credibility of cross-agent collaboration.

[0181] Optionally, in addition to performing the first quantitative assessment and / or the second quantitative assessment as described in S201 and S202 above, the trustworthiness of external resources can also be verified based on a resource call protocol with added identity authentication and call auditing mechanisms.

[0182] External resources include external models, external databases, and external tools.

[0183] For example, an external model refers to a route planning model, traffic flow prediction model, etc., deployed on a third-party server; an external database is an external data storage system that stores regional geographic information, real-time road network status, and resource inventory data; and external tools refer to geographic information query tools, resource scheduling tools, etc., provided by a third party.

[0184] In some embodiments, MCP extends the identity authentication process and the call audit field. The identity authentication adopts a triple authentication mechanism of agent identifier, key and trusted certificate. The call audit field records information such as call time, calling agent identifier, resource usage duration and execution result.

[0185] For example, when a secondary decision-making agent calls an external geographic information system (database resource), it first sends an authentication request to the resource server, carrying its own agent identifier, preset key and trusted certificate issued by the system; after the resource server verifies the authentication, it opens the query interface and records the call audit information (such as call time: XXXX year XX month XX day XX hour XX minute, calling agent: secondary decision-making agent - North Line Transportation No. 1, query content: real-time traffic status of backup road X, execution result: query successful).

[0186] Specifically, during the call, the protocol monitors the data transmission status in real time. If any abnormalities such as data loss or tampering occur, the call is immediately interrupted and an abnormality log is recorded and reported to the execution monitoring agent. After the call ends, the audit information is synchronously uploaded to the system log server for future reference, ensuring that the entire process of resource call is traceable.

[0187] Thus, by extending the security mechanism of the resource invocation protocol, this invention constructs a trusted link for external resource invocation, avoids the risks of untrusted resource access and malicious invocation, and ensures the security of system interaction with external resources.

[0188] The following is combined Figure 3 The complete process of the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by this invention is described in detail.

[0189] For example, such as Figure 3 As shown, the complete process can be divided into the following four main stages:

[0190] Phase 1: Task Initiation and Parallel Planning.

[0191] S301. The commander initiates a task and distributes the task to the superior first-level decision-making intelligent agent.

[0192] For example, the commander defines the core mission objectives based on the actual business scenario requirements (such as emergency support and complex task scheduling), and pushes the mission document to the superior first-level decision-making intelligent agent through a dedicated command terminal as the trigger condition for starting the entire decision-making process.

[0193] S302, The superior first-level decision-making intelligent agent understands the execution intention of the planning intelligent agent and confirms it with the commander.

[0194] For example, after receiving the task document, the superior planning agent analyzes the core objectives, constraints and key requirements of the task, and simultaneously feeds back the intent understanding results to the commander's terminal. After the commander confirms that the intent understanding is accurate, the process proceeds to the next stage.

[0195] S303, The superior authority assesses the execution status of the intelligent agent's plan and confirms it with the commander.

[0196] For example, the superior planning intelligent agent combines the scenario information corresponding to the task (such as the regional road network conditions and resource distribution of the emergency support scenario) to conduct situation analysis, clarify the current task execution difficulties, resource gaps and other information, and submit the judgment results to the commander for confirmation before entering the plan generation stage.

[0197] S304. The superior planning agent generates multiple contingency plans. During this process, the embedded evaluation submodule synchronously performs the prediction evaluation of the contingency plan's adaptability in the first quantitative evaluation. Subsequently, the contingency plans are simulated and deduced, and finally the commander selects and determines the target contingency plan.

[0198] For example, the superior planning agent generates multiple alternative plans based on the confirmed intentions and circumstances; first, the embedded evaluation submodule evaluates the suitability of each plan (such as its fit with the task objectives), and then verifies the feasibility of the plan (such as resource scheduling efficiency) through simulation and deduction. The evaluation and deduction results are then submitted to the commander, who selects the target plan that meets the task requirements.

[0199] S305. The superior planning agent issues a pre-command to the planning agent of the next lower-level decision-making agent.

[0200] For example, the superior planning agent, based on the selected target plan, clarifies the task division and core collaboration requirements of the subordinate agents, and issues pre-commands to the subordinate planning agents.

[0201] S306. The superior planning agent issues its self-generated target plan to the planning agent of the next lower-level decision-making agent.

[0202] S307. Lower-level planning agents, based on pre-issued commands, conduct their own planning work in parallel with higher-level planning agents. The lower-level planning agents understand the execution intent, which is then confirmed by their own commander.

[0203] For example, after receiving a pre-command, the lower-level planning agent parses the task intent and division of labor requirements of the superior, and feeds back its understanding of the intent to the commander at the corresponding level, who then confirms that the understanding is correct.

[0204] S308. The lower-level planning and execution status of the intelligent agent is assessed and confirmed by the commander at this level.

[0205] For example, the lower-level planning agents independently conduct situation analysis based on the actual conditions of their own jurisdiction (such as the resource allocation and geographical environment of the zones), clarify the execution constraints and key links of their own tasks, and submit the judgment results to the corresponding commander for confirmation.

[0206] S309. The lower-level planning agent generates adapted branch plans and conducts simulations; during this process, its embedded evaluation submodule also performs plan adaptability evaluation; finally, the commander at this level selects and determines the target plan at this level.

[0207] For example, the lower-level planning agent generates an appropriate branch plan based on the confirmed intention of the superior and the situation at its own level; it evaluates the plan through its embedded evaluation submodule, verifies its feasibility through simulation and deduction, and submits the results to the corresponding commander to select the target plan at its own level.

[0208] S310. Subordinate planning agents shall report their target plans to their respective levels.

[0209] Phase Two: Resource Preparation and Plan Breakdown.

[0210] S311. The preparation agent of the superior first-level decision-making agent generates a plan based on the target plan. During this process, the evaluation submodule embedded within it synchronously performs the matching evaluation of resource and task matching degree and credibility in the first quantitative evaluation.

[0211] For example, the superior-level intelligent agent, based on the selected target plan, classifies and groups resources such as transportation units, support personnel, and models participating in the task, establishing a clear command hierarchy and organizational collaboration relationships. During this process, the embedded evaluation submodule is invoked to assess the reliability (e.g., equipment availability), controllability (e.g., response speed), and suitability for the task of the resources.

[0212] S312. The superior preparation agent, based on the target plan, organizes resources, establishes command / organization relationships, and executes pre-preparedness.

[0213] For example, the superior preparation agent can take advance actions such as resource scheduling (e.g., equipment maintenance, material loading) and personnel assembly based on the generated plan.

[0214] S313. The superior preparation agent distributes the generated overall plan to the subordinate preparation agents.

[0215] For example, the superior preparation agent will distribute the overall plan, which includes task flow, resource allocation, and time nodes, to the subordinate preparation agent, clarifying the execution requirements and coordination rhythm of the subordinate.

[0216] S314. Preparation of the Lower-Level Decision-Making Agent: Based on the overall plan, the lower-level decision-making agent generates its own plan. Simultaneously, it conducts a matching assessment of resource and task compatibility and reliability.

[0217] For example, the lower-level preparation agent reorganizes the resources and personnel under its jurisdiction according to the overall plan of the higher level, establishes command and cooperation relationships adapted to the tasks of the lower level, and evaluates the matching degree between resources and tasks of the lower level.

[0218] S315. The lower-level prepared agent performs the pre-prepared work of this level.

[0219] For example, a lower-level preparation agent carries out pre-emptive actions such as resource scheduling and personnel preparation based on its own generated branch plan to ensure that the execution rhythm matches that of the higher-level plan.

[0220] S316, The subordinate preparing agent reports the plan it has generated.

[0221] S317. The lower-level preparation agent issues specific branch execution plans to its subordinate secondary decision-making agents.

[0222] For example, the lower-level preparation agent will issue branch plans (such as specific transportation routes and work instructions) refined to the execution level to the corresponding secondary decision-making agent.

[0223] Phase 3: Dynamic execution and monitoring response.

[0224] During this phase, the execution monitoring agents at all levels continuously monitor the task execution status, and their activities include the following parallel or sequential steps:

[0225] S318, The execution monitoring agent of the superior first-level decision-making agent monitors conflicts (such as those related to internal grouping resources).

[0226] For example, the superior execution monitoring agent tracks the resource usage and agent collaboration status in real time during the global task execution process, uses data analysis to identify problems such as resource scheduling conflicts (such as the same device being repeatedly allocated) or agent actions interfering with each other, and records the details of the conflicts.

[0227] S319. The superior execution monitoring agent identifies sudden events (such as external sudden events).

[0228] For example, the superior execution monitoring agent can obtain abnormal information in the task scenario (such as the interruption of major transportation routes in emergency support) through the connected external sensing interface or reports from the subordinate, and immediately conduct an instant assessment of the impact scope, urgency and initial handling direction of the emergency.

[0229] S320: The superior execution monitoring agent evaluates the effectiveness of the event / task execution and executes it after confirmation by the commander.

[0230] For example, the superior execution monitoring agent quantitatively evaluates the progress and phased effects of the current overall task or specific event, and feeds back the evaluation results and suggestions to the commander. After the commander confirms that the execution meets expectations or approves the adjustment plan, the process continues to advance or is adjusted.

[0231] Furthermore, after S320, once the superior monitoring agent receives the application, it immediately calls the embedded evaluation submodule to perform a second quantitative evaluation to determine whether the sudden event has affected / failed to achieve the planned execution effect.

[0232] For example, the monitoring agent obtains information about emergencies through the perception interface and calls the embedded evaluation submodule to conduct real-time evaluation of the scope of the event's impact, the difficulty of handling it, and its effect on the overall plan.

[0233] S321. If it is determined that the plan has been affected or has not achieved its intended effect, then the plan should be adjusted.

[0234] S322, The execution monitoring agent of the lower-level first-level decision-making agent monitors conflicts (such as internal grouping resources).

[0235] For example, the lower-level execution monitoring agent focuses on the execution of its own branch plan, monitors the use of resources and the collaboration status of the agent, identifies and records resource conflicts or operational conflicts within its jurisdiction, and simultaneously reports important conflict information to the upper-level execution monitoring agent.

[0236] S323, The lower-level execution monitoring agent identifies sudden events (such as external sudden events).

[0237] For example, when a lower-level execution monitoring agent detects an anomaly in its area of ​​responsibility (such as a designated transport convoy being unable to pass as planned), it immediately conducts a rapid assessment of the impact and difficulty of handling the local emergency, and pushes the event information and preliminary assessment results to the higher-level execution monitoring agent.

[0238] S324. The subordinate execution monitoring agent evaluates the effect of the event / task execution and executes it after confirmation by the commander.

[0239] For example, the lower-level execution monitoring agent continuously evaluates the execution effect of events or tasks at its level, and feeds back the evaluation results to the commander at its level or the commander at the higher level. After confirmation by the commander, the agent guides the execution or adjustment of subsequent actions at its level.

[0240] Furthermore, after S324 above, after the lower-level execution monitoring agent receives the application, it immediately calls the embedded evaluation submodule to perform a second quantitative evaluation to determine whether the sudden event affects / fails to achieve the planned execution effect.

[0241] For example, the monitoring agent obtains information about emergencies through the perception interface and calls the embedded evaluation submodule to conduct real-time evaluation of the scope of the event's impact, the difficulty of handling it, and its effect on the overall plan.

[0242] S325. If it is determined that the plan has been affected or has not achieved its intended effect, the plan shall be adjusted.

[0243] In the context of this continuous monitoring, emergencies may be detected by the secondary decision-making agent and trigger the following dynamic response closed loop.

[0244] S326. The secondary decision-making agent identifies unexpected events during the execution of the plan.

[0245] For example, during the execution of a branch plan, the secondary decision-making agent obtains abnormal information in the task scenario (such as road collapse during material transportation) through its own perception module and identifies sudden events that exceed the normal execution scope.

[0246] S327. The secondary decision-making agent assesses the event and determines that it cannot be handled independently.

[0247] For example, a secondary decision-making agent assesses the difficulty of handling an emergency based on its own execution authority, resource allocation, and handling capabilities, and determines whether it has the conditions to handle the situation independently.

[0248] S328. If it is determined that the task cannot be completed independently, the secondary decision-making agent initiates an empowerment request to the lower-level execution monitoring agent, which then forwards the request to the higher-level execution monitoring agent.

[0249] For example, when the secondary decision-making agent determines that it cannot handle the situation independently, it sends an empowerment request to the superior execution monitoring agent.

[0250] S329. After authorization by the commander, the secondary decision-making agent that submitted the application is authorized and its role identifier is updated and / or it is empowered.

[0251] For example, after being authorized by the commander, the secondary decision-making agent obtains temporary command and decision-making authority and / or supplementary resource and capability interfaces, and after obtaining temporary command and decision-making authority, updates its role identifier to a temporary primary decision-making agent.

[0252] S330, the secondary decision-making agent, parses and generates disposal instructions.

[0253] It also analyzes and handles emergency scenarios to generate specific execution instructions (such as scheduling backup resources and adjusting transportation routes).

[0254] S331, The secondary decision-making intelligent agent executes the disposal instructions.

[0255] For example, the temporary first-level decision-making agent, based on the generated handling instructions, calls upon the corresponding collaborating agents and resources to execute the handling operations for the emergency.

[0256] S332, The secondary decision-making agent judges the task completion status.

[0257] For example, a temporary first-level decision-making agent evaluates the progress and effectiveness of the handling operation.

[0258] S333. If the plan requirements are not met, continue execution.

[0259] For example, if the evaluation results show that the disposal objective has not been achieved, the temporary first-level decision-making agent adjusts the execution strategy and continues to execute the disposal instructions.

[0260] Furthermore, if the plan requirements are met, the judgment steps following S324 above will be executed.

[0261] Furthermore, after the temporary Level 1 decision-making agent has completed its task, its permissions are released, and its role is restored to that of a Level 2 decision-making agent.

[0262] For example, once the emergency response is completed and the desired effect is achieved, the system automatically revokes the temporary permissions and interfaces configured for the agent and restores its role identifier to the original level-two decision-making agent.

[0263] Phase Four: Overall Effectiveness Evaluation and Summary

[0264] S334. If it is determined after S320 that the sudden event has not affected / achieved the planned execution effect, then the independent evaluation agent (second working mode) of the superior first-level decision-making agent collects and summarizes the data.

[0265] For example, the superior evaluation agent collects data from the entire task execution process (such as contingency plan execution data, resource usage data, and emergency response data), and then categorizes, summarizes, and structures the data.

[0266] S335. The superior evaluation agent performs the third quantitative evaluation to assess the overall effectiveness of the decision-making plan, which is then confirmed by the commander.

[0267] For example, the superior evaluation agent conducts a third quantitative evaluation of the overall plan's execution effect and goal achievement based on the aggregated data, and submits the evaluation results to the commander for confirmation.

[0268] S336. If the desired effect is achieved, generate a final summary report.

[0269] For example, if the evaluation results show that the implementation of the plan meets the standards, the superior evaluation agent generates a task summary report, which becomes the final document after being confirmed by the commander.

[0270] S337. If the desired effect is not achieved, adjust the plan and re-execute it.

[0271] S338. If, after S324, it is determined that the sudden event did not affect / achieve the planned execution effect, then the independent evaluation agent of the next-level decision-making agent collects and summarizes the data.

[0272] For example, the lower-level evaluation agent collects various types of data during the execution of its own tasks, categorizes and summarizes them, and then synchronizes them to the upper-level evaluation agent.

[0273] S339. The lower-level evaluation agent evaluates the overall completion effect of the plan at this level, and the evaluation is confirmed by the commander.

[0274] For example, the lower-level evaluation agent evaluates the execution effect of the branch plan based on the aggregated data at its level, and submits the results to the corresponding commander for confirmation.

[0275] S340. If the desired effect is achieved, generate a summary report at this level and synchronize it to the superior.

[0276] For example, if the execution of the plan at this level meets the target, the lower-level evaluation agent generates a task summary report for this level, which, after being confirmed by the corresponding commander, is synchronized to the higher-level evaluation agent.

[0277] S341. If the desired effect is not achieved, adjust the plan and re-execute it.

[0278] For example, if the evaluation results show that the planned effect has not been achieved, the lower-level evaluation agent adjusts the branch plan based on the cause of the problem, and after confirmation by the corresponding commander, it is reissued for execution.

[0279] Thus, this invention achieves efficient collaboration, dynamic adaptation, and reliable controllability in swarm intelligence decision-making through a hierarchical parallel and dynamically responsive intelligent agent collaboration process, as well as a quantitative evaluation mechanism throughout the process, effectively improving the flexibility, reliability, and traceability of task execution in complex scenarios.

[0280] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communications bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other through the communications bus 440.

[0281] The processor 410 can call logical instructions in the memory 430 to execute a swarm intelligence-based trusted human-machine hybrid intelligent decision-making method. This method includes: generating a decision scheme by a first-level decision agent and distributing an execution plan derived from the decomposition of the decision scheme to a second-level decision agent; wherein the process of generating the decision scheme includes a first quantitative assessment, which at least includes a predictive assessment of the suitability of the plan and a matching assessment of the credibility and controllability of the resources involved in the decision scheme; receiving an empowerment request initiated by a second-level decision agent during the execution of the execution plan after it has autonomously determined that the unforeseen event cannot be handled independently, and performing a second quantitative assessment. The assessment should include at least an evaluation of the impact of handling emergencies on the effectiveness of plan execution; in response to an empowerment request and with the authorization of the commander, dynamically configure temporary command and decision-making permissions and / or supplementary resource and capability interfaces for the secondary decision-making agent that submitted the empowerment request, and update its role identifier after configuring the temporary command and decision-making permissions; after the emergency is handled, reclaim the temporary command and decision-making permissions and supplementary resource and capability interfaces dynamically configured for the secondary decision-making agent, and restore its role identifier to its original state; monitor the assessment results of the first or second quantitative assessment, and when the assessment result is lower than the corresponding threshold, interrupt or suspend the current process and request the commander's intervention.

[0282] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0283] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by the above methods. The method includes: generating a decision scheme by a first-level decision intelligence agent and distributing the execution plan obtained from the decomposition of the decision scheme to a second-level decision intelligence agent; wherein, the process of generating the decision scheme includes a first quantitative evaluation, which includes at least a predictive evaluation of the suitability of the scheme and a matching evaluation of the credibility and controllability of the resources involved in the decision scheme; receiving the second-level decision intelligence agent's autonomous judgment that it cannot independently execute the execution plan during the execution process. The system initiates an empowerment request after handling an emergency and performs a second quantitative assessment, which includes at least an assessment of the impact of the emergency handling on the effectiveness of the plan execution. In response to the empowerment request and with the commander's authorization, it dynamically configures temporary command and decision-making permissions and / or supplementary resource and capability interfaces for the secondary decision-making agent that submitted the empowerment request, and updates its role identifier after configuring the temporary command and decision-making permissions. After the emergency is handled, it reclaims the temporarily configured command and decision-making permissions and supplementary resource and capability interfaces for the secondary decision-making agent and restores its role identifier to its original state. It monitors the assessment results of the first or second quantitative assessment; when the assessment result is lower than the corresponding threshold, it interrupts or suspends the current process and requests the commander's intervention.

[0284] Furthermore, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the trusted human-machine hybrid intelligent decision-making method based on swarm intelligence provided by the above methods. This method includes: generating a decision scheme by a first-level decision intelligence agent and distributing an execution plan obtained from the decomposition of the decision scheme to a second-level decision intelligence agent; wherein the process of generating the decision scheme includes a first quantitative assessment, which at least includes a predictive assessment of the suitability of the plan and a matching assessment of the credibility and controllability of the resources involved in the decision scheme; and receiving an empowerment initiated by the second-level decision intelligence agent during the execution of the execution plan after it autonomously determines that the unforeseen event cannot be handled independently. The system applies for and executes a second quantitative assessment, which includes at least an assessment of the impact of the emergency response on the effectiveness of the plan execution. In response to the empowerment application and with the commander's authorization, it dynamically configures temporary command and decision-making permissions and / or supplementary resource and capability interfaces for the secondary decision-making agent that submitted the empowerment application, and updates its role identifier after configuring the temporary command and decision-making permissions. After the emergency response is completed, it reclaims the temporarily configured command and decision-making permissions and supplementary resource and capability interfaces for the secondary decision-making agent and restores its role identifier to its original state. It monitors the assessment results of the first or second quantitative assessment; when the assessment result is lower than the corresponding threshold, it interrupts or suspends the current process and requests the commander's intervention.

[0285] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0286] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0287] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents, characterized in that, The method is executed based on a hierarchical cascaded architecture comprising a first-level decision agent and a second-level decision agent; the method includes: The first-level decision-making agent generates a decision scheme and distributes the execution plan obtained from the decomposition of the decision scheme to the second-level decision-making agent. The process of generating the decision scheme includes a first quantitative evaluation, which includes at least predictive evaluation of the adaptability of each set of contingency plans and matching evaluation of the credibility and controllability of the resources involved in each set of contingency plans. The decision scheme is determined from the multiple sets of contingency plans. The system receives an empowerment request initiated by the secondary decision-making agent during the execution of the execution plan after the agent autonomously determines that the emergency is one that cannot be handled independently. The system then performs a second quantitative assessment, which includes at least an assessment of the impact of the proposed handling plan on the emergency and its impact on the execution effect of the plan. In response to the empowerment request and with the authorization of the commander, the secondary decision-making agent that made the empowerment request is dynamically configured with temporary command and decision-making permissions and / or supplementary resource and capability interfaces, and its role identifier is updated after configuring the temporary command and decision-making permissions. After the emergency is handled, the temporary command and decision-making authority and the supplementary resource and capability interfaces that were dynamically configured for the secondary decision-making agent are revoked, and its role identifier is restored to its original state. Monitor the evaluation results of the first or second quantitative assessment. When the evaluation result is lower than the corresponding threshold, interrupt or suspend the current process and request the commander to intervene.

2. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 1, wherein the first-level decision-making intelligence agent includes a planning intelligence agent, a preparation intelligence agent, an execution monitoring intelligence agent, and an evaluation intelligence agent; in, The first-level decision-making agents at the same level support sequential collaboration in a ring-shaped cascade, and the first-level decision-making agents at different levels support cross-level vertical linkage collaboration. The secondary decision-making agent is used for: Execute the execution plan issued by the first-level decision-making agent; or, When an emergency that cannot be handled independently is detected, an empowerment request is initiated to obtain the temporary command and decision-making authority and / or the supplementary resource and capability interfaces required for parallel and collaborative execution with other secondary decision-making agents at the same level. or, After obtaining the temporary command and decision-making authority and updating the role identifier to a temporary level-one decision-making intelligent agent, organize other intelligent agents to coordinate the handling of the emergency.

3. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 2, characterized in that, The evaluation agent has a first working mode and a second working mode; The first working mode is an evaluation submodule, which is embedded in the planning agent, the preparation agent and the execution monitoring agent respectively; The second working mode is an independent evaluation agent.

4. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 3, characterized in that, An evaluation submodule embedded in the planning agent is used to perform a predictive evaluation of the adaptability of each set of plans in the first quantitative evaluation. The evaluation submodule embedded in the preparation agent is used to perform the matching evaluation of the credibility, controllability and task execution controllability of the resources involved in each plan in the first quantitative evaluation. An evaluation submodule embedded in the execution monitoring agent is used to perform the second quantitative evaluation of the impact of the proposed handling plan for the emergency and its impact on the execution effect of the plan; The independent evaluation agent is used to perform a third quantitative evaluation of the overall effectiveness of the decision-making scheme.

5. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 2, characterized in that, The planning agent includes a superior planning agent and subordinate planning agents. The steps for generating the decision scheme include: The higher-level planning agent receives the task and issues advance orders to the lower-level planning agent based on the commander's intentions; Multiple contingency plans are generated in parallel by the superior planning agent and the subordinate planning agent. Each contingency plan is simulated and analyzed, and the decision-making scheme is determined according to the commander's instructions.

6. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 2, characterized in that, The preparation agent includes a superior preparation agent and a subordinate preparation agent. The steps for generating the execution plan include: The superior preparation agent organizes and evaluates resources according to the decision-making scheme, generates a local execution plan, and issues it. The lower-level preparation agent receives the execution plan of this level, generates an adapted branch plan, and sends it to the corresponding secondary decision-making agent.

7. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 2, characterized in that, The step of dynamically configuring the secondary decision agent that submits the empowerment application includes: The execution monitoring agent evaluates the empowerment application and generates an authorization proposal. Upon receiving confirmation from the commander regarding the authorization proposal, the temporary command and decision-making authority and / or the supplementary resource and capability interfaces are dynamically configured for the secondary decision-making agent that submitted the empowerment application. After configuring the temporary command and decision-making permissions, update its role identifier.

8. The trusted human-machine hybrid intelligent decision-making method based on swarm intelligence agents according to claim 1, characterized in that, The method further includes: Based on an inter-agent communication protocol that incorporates trusted and controllable value information fields, the collaborative state of the agents is monitored and interactively verified. Based on a resource call protocol with added identity authentication and call auditing mechanisms, the trustworthiness of external resources is verified; The external resources include external models, external databases, and external tools.

9. A trusted human-machine hybrid intelligent decision-making system based on swarm intelligence, characterized in that, The system includes: A first-level decision agent is configured to perform the operations performed by the first-level decision agent in the method according to any one of claims 1-8; A secondary decision agent is configured to perform the operations executed by the secondary decision agent in the method according to any one of claims 1-8; The Trusted and Controllable Module is deployed among various intelligent agents and is configured to monitor and interactively verify the collaborative state of intelligent agents based on an inter-agent communication protocol with added trusted and controllable value information fields. The resource management module, deployed between the intelligent agent and external resources, is configured to verify the trustworthiness of external resources based on a resource call protocol with added identity authentication and call auditing mechanisms. The external resources include external models, external databases, and external tools.

10. The trusted human-machine hybrid intelligent decision-making system based on swarm intelligence as described in claim 9, characterized in that, The first-level decision-making agent includes a planning agent, a preparation agent, an execution monitoring agent, and an evaluation agent. The evaluation agent has a first working mode and a second working mode; The first working mode is an evaluation submodule, which is embedded in the planning agent, the preparation agent and the execution monitoring agent respectively; The second working mode is an independent evaluation agent; Among them, sequential collaboration in a ring-shaped cascade is supported between the first-level decision-making agents at the same level, and cross-level vertical linkage collaboration is supported between the first-level decision-making agents at different levels. The secondary decision-making agent is used for: Execute the execution plan issued by the first-level decision-making agent; or, When an emergency that cannot be handled independently is detected, an empowerment request is initiated to obtain the temporary command and decision-making authority and / or the supplementary resource and capability interfaces required for parallel and collaborative execution with other secondary decision-making agents at the same level. or, After obtaining the temporary command and decision-making authority and updating the role identifier to a temporary level-one decision-making intelligent agent, organize other intelligent agents to coordinate the handling of the emergency.

Images