Meteorological service system based on multiple agents and system construction and optimization method

Abstract

The invention discloses a multi-agent-based meteorological service system and a system construction and optimization method, and the system comprises an interaction agent which is used for regulating and controlling the whole system and achieving the barrier-free communication between a user and each part of the system; the meteorological agent is used for querying the meteorological information and other meteorological information related to user activities; the planning agent is used for carrying out modeling and optimization of activity time and space and carrying out user activity simulation optimization with the behavior agent; and the behavior agent is used for simulating behavior logic of the user in a personalized manner and giving an optimization request with preference and pertinence. According to the invention, modeling is carried out on the weather information related to the user activity, and personalized correction is carried out by simulating the user, so that the weather-oriented activity arrangement of the user is improved, the experience of the weather service is enhanced, the weather-oriented activity arrangement is visually visualized, and the user experience is improved. And a more personalized and perfect weather management scheme in an activity can be provided for the user.

Description

Technical Field

[0001] This invention belongs to the field of meteorological service technology, and more specifically, relates to a meteorological service system based on multiple agents and a method for system construction and optimization. Background Technology

[0002] Consideration of meteorological factors is crucial in daily activities, directly determining the success and effectiveness of events, especially under complex spatiotemporal conditions. Current meteorological services for individual users often focus on two aspects: first, providing real-time information on potential abnormal weather conditions for the current location and the next few hours, alerting users to take precautions and avoid potential risks; second, providing a broader forecast of future weather conditions for a specified area, allowing users to independently query abnormal weather conditions at any location that might hinder their activities. Traditional meteorological services are limited by query methods and interaction modes, confined to real-time or on-site conditions, making it difficult to provide users with more personalized and comprehensive meteorological management solutions for their activities. The user experience when using a meteorological service system is crucial; that is, traditional meteorological service systems lack deep understanding and feedback capabilities regarding user activities. Therefore, this invention provides a multi-agent-based meteorological service system and a method for its construction and optimization. Summary of the Invention

[0003] This invention addresses the technical problems existing in the prior art by providing a meteorological service system based on multiple agents and a method for system construction and optimization. The system can optimize the user experience when using meteorological services to plan long-term activities, and the method can assist users in planning daily activities based on multiple agents to avoid the impact of abnormal weather conditions.

[0004] According to a second aspect of the present invention, the present invention provides a multi-agent-based meteorological service system, comprising: Interactive agents are used to interact with users and other agents, convey information, clarify needs, and act as intermediaries to enable other agents to effectively understand user needs, while enabling users to quickly understand the weather-based activity plan given by the system as a whole and provide feedback. Meteorological intelligent agents are used to query and organize meteorological information, including professional and accurate real-time meteorological information and spatiotemporal information of events; The planning agent is used to plan the activities provided by the user in the time and space, and optimize the activity time and space by combining the information and feedback provided by the meteorological agent and the behavioral agent. After the target conditions are met and the maximum number of iterations is reached, the designed activity time and space is output. Behavioral intelligence agents are used to personalize and simulate user behavior logic and provide optimization requests with preferences and specificity, including: simulating user activity planning, using the activity space-time established by the planning intelligence agent as the activity range, experiencing and evaluating the planning in the activity space, and providing nodes to be optimized for iterative optimization.

[0005] Based on the above technical solution, the present invention can also be improved as follows.

[0006] Preferably, the interactive intelligent agent interacts with the user and other intelligent agents, conveying information including: user needs, user-specified modification methods, and the user profile reflected by the user; this information is then exchanged with the meteorological intelligent agent, the planning intelligent agent, the behavioral intelligent agent, and local storage, respectively; the feedback from the meteorological intelligent agent and the planning intelligent agent on the needs is refined by the interactive intelligent agent and fed back to the user; the interactive intelligent agent performs global planning of the system, providing guidance for the overall system.

[0007] Preferably, the professional and accurate real-time meteorological information and activity spatiotemporal information are obtained through authoritative third-party meteorological agencies, and the activity spatiotemporal information is obtained by filtering from knowledge bases and external Internet searches, or provided by users; the meteorological agent can query authoritative meteorological information provided by credible third-party authoritative meteorological service providers, and can search for meteorological guidance on relevant activities in the activity area through official media, self-media, and social platforms on search engines. At the same time, based on the feedback from the planning agent, it can query more detailed meteorological information, providing the planning agent with global information reference and unofficial information.

[0008] Preferably, the planning agent can combine all the received information to optimize and construct the activities proposed by the user, provide a complete solution, and make comprehensive plans based on user experience and personalized preferences.

[0009] Preferably, the behavioral intelligence agent can simulate the user's behavioral logic and perform personalized construction on the basis of conventional logic, including: simulating conventional user-defined behavioral logic in the state of not receiving a user profile, and reading local storage and modifying the predefined user profile during user use so that the behavioral logic conforms to user preferences.

[0010] According to a second aspect of the present invention, a method for constructing and optimizing a multi-agent-based meteorological service system is disclosed, for use in the aforementioned system, the method comprising the following steps: The interactive intelligent agent interacts with the user to obtain the activities the user needs to perform, and combines the Internet to obtain weather information related to the activities; The meteorological intelligent agent performs queries on user activities and the spatiotemporal meteorological environment; The planning and construction activities of the intelligent agent are planned in time and space, simulating the behavioral logic of users in specific activities, and combining the personal information provided by users to carry out personalized planning and modeling. The behavioral agent simulates activities according to the pre-defined activity plan time and space, provides real-time feedback on the problems existing in the current activity time and space, and the interactive agent receives and organizes the feedback and provides feedback to the user. The planning agent processes the feedback and feeds the results back to the planning agent for iterative optimization. The iterative process repeats until the behavioral agent completes the full activity simulation. The planning agent visualizes the activity's time and space, and then hands it over to the interactive agent for user display.

[0011] Based on the above technical solution, the present invention can also be improved as follows.

[0012] Preferably, the activities that the user needs to perform include: temperature, humidity, wind speed, meteorological conditions and corresponding sensory information during the activity time; the meteorological information related to the activity obtained by combining the Internet includes: the specific content of the activity, the duration of the activity, and meteorological precautions related to the activity; the meteorological agent's query of the user's activities and the spatiotemporal meteorological environment includes: the expected spatiotemporal location of the interactive agent and the activity execution, the dynamic spatiotemporal transition between activities, a spatiotemporal location search performed once every preset time period, and the detection of user status and meteorological environment in subsequent spatiotemporal periods.

[0013] Preferably, the activity simulation includes: states at activity points and movement between different activity points; the behavioral agent performs activity simulation according to a preset activity planning time and space, including: The behavioral agent and the planning agent interact iteratively, and the spatiotemporal activities of the meteorological domain planned by the planning agent are evaluated and optimized through real-time feedback and tolerance accumulation. The evaluation and optimization formula is expressed as follows:

[0014] in, This refers to the cumulative tolerance level over time. To describe the inappropriateness of the current meteorological environment, t represents time. A negative constant indicates that the cumulative tolerance decreases over time; by summing the tolerance at multiple time points, the total tolerance becomes... If the user exceeds the tolerance threshold, it is considered that the user cannot tolerate the discomfort, the simulation is terminated, and feedback is sent to the planning agent.

[0015] Preferably, the planning agent further processes the feedback by: If the planning agent does not process the feedback or accepts the problem, the current activity spacetime is maintained, and the simulation of the behavioral agent is not stopped until the end of the current activity simulation; if the user proposes modifications, the activity spacetime is completely reset or replanned from the previous spacetime node.

[0016] Preferably, the planning agent visualizes and outputs the activity spatiotemporal information, including: The planning agent divides the map information of the activity space into multi-level map formats according to the size of the activity space, and specifies interactive design for activity spatiotemporal points. When interacting with activity points, it displays the overall meteorological information of the current map in the form of a meteorological map. The interactive content includes: abnormal meteorological events encountered and meteorological indicators affecting the activity. Compared with existing technical solutions, the present invention has the following significant advantages: This invention provides a meteorological service system based on multi-agent intelligence and a method for system construction and optimization. First, by acquiring user information and additional external meteorological service information, the multi-agent meteorological service system realizes the modeling and optimization of activity spatiotemporal information, improves the user experience when querying meteorological services, and ensures the efficiency and reliability of activity spatiotemporal information. Secondly, the multi-agent system built on the CrewAI framework of this invention has good multi-agent communication efficiency and cooperation capabilities. Through practice, the division of labor among the multi-agents is carefully carried out to avoid inefficiency and cost increase caused by invalid communication between multi-agents, effectively accelerating the activity planning speed and improving the user experience. Finally, this invention achieves the planning and optimization of meteorological aspects of user activities through activity spatiotemporal modeling and spatiotemporal node experience. Users can preview and further plan the overall activity through the provided interactive activity spatiotemporal route map, thereby improving the user experience of the meteorological service system. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the structure of a multi-agent-based meteorological service system provided in an embodiment of the present invention; Figure 2 A flowchart illustrating the construction and optimization method of a meteorological service multi-agent system provided in this embodiment of the invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Understandably, given the deficiencies in the prior art, according to a first aspect of the present invention, the present invention provides a multi-agent-based meteorological service optimization system, specifically as follows: Figure 1 As shown, it includes: Interactive agents are used to interact with users and other agents, convey information, clarify needs, and act as intermediaries to enable other agents to effectively understand user needs, while enabling users to quickly understand the weather-based activity plan given by the system as a whole and provide feedback. Meteorological intelligent agents are used to query and organize meteorological information, including professional and accurate real-time meteorological information and spatiotemporal information of events; The planning agent is used to plan the activities provided by the user in the time and space, and optimize the activity time and space by combining the information and feedback provided by the meteorological agent and the behavioral agent. After the target conditions are met and the maximum number of iterations is reached, the designed activity time and space is output. Behavioral intelligence agents are used to personalize and simulate user behavior logic and provide optimization requests with preferences and specificity, including: simulating user activity planning, using the activity space-time established by the planning intelligence agent as the activity range, experiencing and evaluating the planning in the activity space, and providing nodes to be optimized for iterative optimization.

[0020] Specifically, the interactive intelligent agent interacts with the user and other intelligent agents to convey necessary information, including user needs, user-specified modification methods, and the user profile. This information will be exchanged with the meteorological intelligent agent, the planning intelligent agent, the behavioral intelligent agent, and local storage, respectively. At the same time, the feedback from the meteorological intelligent agent and the planning intelligent agent on the needs will also be refined through the interactive intelligent agent and fed back to the user. In addition, the interactive intelligent agent also performs global planning of the system, providing guidance for the overall system and preventing communication efficiency reduction and abnormal intelligent agent startup caused by systemic structural issues, including excessive task execution and infinite optimization iteration of the planning intelligent agent and the meteorological intelligent agent. The meteorological agent can query authoritative meteorological information provided by credible third-party authoritative meteorological service providers, and can search for meteorological guidance for relevant activities in the activity area on relevant official media, self-media, social platforms, etc. on search engines. At the same time, it can query more meteorological-related detailed information based on the feedback of the planning agent, providing the planning agent with global information reference and unofficial information, and comprehensively improving the user's overall experience. The planning agent can combine all the information received to optimize and construct the activities proposed by the user and provide a complete solution. The feature of this agent is that the planning process combines actual time and user profile, from the overall layout to the implementation of specific details, and can carry out comprehensive planning based on user experience and personalized preferences. The behavioral agent simulates the user's behavioral logic and can be personalized on the basis of conventional logic. Specifically, the behavioral agent simulates conventional user-defined behavioral logic when it has not received a user profile. During the user's use, the interactive agent stores the user profile locally, and the behavioral agent reads the local storage and modifies the predefined user profile to make the behavioral logic more in line with the user's preferences, rather than popular preferences.

[0021] The system includes local storage for storing local knowledge bases and user profiles.

[0022] It should be noted that the multi-agent-based meteorological service system described in this embodiment of the invention is suitable for supplementing users' incomplete meteorological activity plans and evaluating and optimizing existing current meteorological planning schemes. The specific process is as follows: For incomplete activity plans, the interactive agent tags the plan and identifies its completion status. Using existing information as a reference, it prompts the planning agent to complete an initial full plan for the activity and hand it over to the interactive agent. The interactive agent then transcribes the full plan into text, provides it to the user, and awaits the next action. Next, it enters the function of evaluating existing complete weather planning schemes. In this process, the interactive agent directly transforms the complete plan into an external activity space and hands it over to the action agent, skipping the planning and weather agents. The action agent independently performs simulations until a tolerance threshold is reached or the simulation is completed, and transmits the saved historical state information to the interactive agent. The interactive agent analyzes the historical state information of the action agent and the corresponding activity spatiotemporal points, and reports the evaluation results of the provided full plan to the user in plain language, providing possible preferred optimization schemes.

[0023] In summary, the multi-agent system built on the CrewAI framework of this invention has good multi-agent communication efficiency and collaboration capabilities. Through careful division of labor among the multi-agents in practice, it avoids inefficiency and cost increases caused by ineffective communication between multi-agents, effectively accelerates activity planning speed, and improves the user experience.

[0024] According to a second aspect of the present invention, embodiments of the present invention propose a method for constructing and optimizing a meteorological service system based on multi-agent systems, specifically as follows: Figure 2 As shown, the method includes the following steps: S1: The interactive intelligent agent interacts with the user to obtain the content of the activities the user needs to perform, and combines the Internet to obtain weather information related to the activities; Understandably, the activity content includes the temperature, humidity, wind speed, meteorological conditions and corresponding sensory information during the activity time; the meteorological information related to the activity obtained by combining the Internet includes: the specific content of the activity, the duration of the activity, and meteorological precautions related to the activity. The multi-agent environment is initialized. When a user actively wakes up the weather service system and provides specific activity content and time requirements, the interactive agent first uses the underlying large model to understand the specific content of the user input and decomposes it into subtasks, breaking down a single complex task into multiple simple subtasks. For the decomposed subtasks, the weather agent searches the Internet step by step to obtain the meteorological information necessary for activity planning. Understandably, the active wake-up method can be audio commands, text commands, etc., and there are no restrictions here.

[0025] Understandably, during the task decomposition process, the interactive intelligent agent divides known tasks into different sub-tasks based on function and complexity. Specifically, tasks requiring the retrieval of activity-related meteorological information, such as activity time and location, are assigned as sub-tasks to the meteorological intelligent agent for information retrieval. Sub-tasks requiring no additional third-party meteorological information, such as the activity's inherent attributes, duration, and meteorological requirements, are directly assigned to the planning intelligent agent. The overall activity plan is then planned after the meteorological intelligent agent completes the meteorological information retrieval for its sub-task and submits it to the planning intelligent agent. Understandably, during the step-by-step search process, the meteorological intelligent agent queries the specific time and region involved in the user's activity and uses search engines and third-party services, primarily referring to APIs provided by search engines and map service providers, as well as large model knowledge bases, to retrieve activity-related meteorological information.

[0026] The system predicts the spatiotemporal locations of interactive intelligent agents and activities, dynamically manages the spatiotemporal transitions between activities, performs a spatiotemporal location search every 10 minutes, and detects user status and weather conditions in subsequent spatiotemporal events.

[0027] S2: The meteorological agent performs queries on user activities and the spatiotemporal meteorological environment; The meteorological agent can query authoritative meteorological information provided by credible third-party authoritative meteorological service providers, and can search for meteorological guidance for relevant activities in the activity area on relevant official media, self-media, social platforms, etc. on search engines. At the same time, it can query more detailed meteorological information based on the feedback of the planning agent, providing the planning agent with global information reference and unofficial information, and comprehensively improving the user's overall experience.

[0028] The activity content includes the temperature, humidity, wind speed, meteorological conditions and corresponding sensory information during the activity time; the meteorological information related to the activity obtained by combining the Internet includes: the specific content of the activity, the duration of the activity, and meteorological precautions related to the activity. The meteorological agent performs user activity and spatiotemporal meteorological environment queries, including: The system includes the predicted spatiotemporal locations of interactive intelligent agents and activities, the dynamic spatiotemporal transitions between activities, a spatiotemporal location search performed once every preset time period, and the detection of user status and weather conditions in subsequent spatiotemporal events.

[0029] S3: The planning agent constructs activity spatiotemporal information related to weather conditions, uses a large model to simulate the behavioral logic of real user behavior, and combines the personal information provided by the user to carry out personalized planning. Understandably, the planning agent receives activity information and activity-related sub-tasks provided by the meteorological agent and the interaction agent, and designs activity spatiotemporal and action plans for the behavioral agent. The activity spatiotemporal is optimized by the locations involved in the actual activities, and the activity spatiotemporal points in the activity space are configured using path planning provided by a third-party map service provider.

[0030] Understandably, activity spatiotemporal optimization design refers to selecting the action path with the shortest route / travel time while meeting the user's needs for the activity.

[0031] Without providing additional activity-related information, the planning agent automatically arranges the optimal travel mode and route for the user that meets the conditions. For activities that require carrying a lot of items, the default is to use motor vehicles as the means of transportation rather than public transportation, and the activity's spatiotemporal points are planned and configured along the path with the shortest arrival time.

[0032] The planning agent utilizes a large model to construct a behavioral agent capable of simulating real user behavior logic, satisfying the following conditions: The behavioral logic of the intelligent agent should match that of a regular user, and it should be able to identify user preferences and needs proactively provided by the user, overlaying relevant information onto the original basic user profile. Simultaneously, it should be able to respond to corresponding activity points in time and space based on the new behavioral logic, offering relevant opinions and suggestions. Furthermore, the intelligent agent should be able to save and retrieve the state after each activity point ends, accumulating states to ensure that long-term low-impact meteorological conditions are not ignored as noise. Understandably, the aforementioned "regular user" refers to someone with basic awareness of temperature and meteorological conditions, and possesses fundamental problem-solving abilities.

[0033] When the planning agent receives historical feedback from the behavioral agent, it analyzes each major activity spatiotemporal point in the history and provides feedback to the user through the interactive agent. If the user does not provide further feedback or the response times out, the planning agent optimizes the activity spatiotemporal on its own and interacts with the behavioral agent to iteratively optimize the activity spatiotemporal.

[0034] It should be noted that the optimization described in this embodiment of the invention can be divided into two parts: experience optimization and path optimization. Experience optimization refers to the additional consideration of the cumulative impact of minor meteorological conditions, which include minor temperature anomalies, minor humidity anomalies, drizzle, minor direct sunlight, and prolonged poor road conditions. At the same time, during the activity, local storage and Internet information, such as social networks and other media, self-media, etc., are combined with the known conditions of the activity to evaluate the state of the activity. Experience optimization is an additional prerequisite for path optimization. It considers the conventional conditions and experience optimization conditions to model and select a spatiotemporal space that is more in line with personal characteristics.

[0035] S4: The behavioral agent simulates activities according to a preset activity time and space, including the state at each activity point and the movement between different activity points; and provides real-time feedback on problems existing in the current activity time and space, which are received and organized by the interactive agent and then fed back to the user. After the behavioral agent is constructed and can execute the specific tasks proposed by the user, that is, simulate the real user's real feelings and feedback on the meteorological environment, the planning agent and the behavioral agent interact multiple times to simulate each activity spatiotemporal point in chronological order. When the behavioral agent receives the information of the current activity spatiotemporal point, it simulates the current state based on the historical state and user profile, and stops the activity after reaching the tolerance threshold or completing the full simulation, and feeds back the historical state of each activity spatiotemporal point to the planning agent.

[0036] Understandably, the activity spatiotemporal points are sampled every 10 minutes. The spatial location is obtained by the optimal path planning simulation provided by the map API, and the corresponding meteorological information is obtained using this time and spatial information.

[0037] Understandably, the initialization of the behavioral agent uses a public overall user profile, and the encrypted information stored in the local preference library during use will affect the subsequent construction of the behavioral agent.

[0038] The final historical state feedback is relayed back to the planning agent in a chain. The feedback uses a one-time format check. If any field is missing, the chain is considered invalid and a retransmission is automatically requested. The maximum number of retries is 3. If the number of retries is exceeded, the interactive agent is used to search for historical information, generate historical state feedback, and check the format again. The maximum number of retries is 3. If an error is found, the current round of operation is terminated.

[0039] The behavioral agent and the planning agent interact iteratively, and the spatiotemporal activities of the meteorological domain planned by the planning agent are evaluated and optimized through real-time feedback and tolerance accumulation. The evaluation and optimization formula is expressed as follows:

[0040] in, This refers to the cumulative tolerance level over time. To describe the inappropriateness of the current meteorological environment, t represents time. A negative constant indicates that the cumulative tolerance decreases over time. By summing the tolerance values ​​at multiple time points, the total tolerance is calculated. If the user exceeds the tolerance threshold, it is considered that the user cannot tolerate the discomfort, the simulation is terminated, and feedback is sent to the planning agent.

[0041] Understandably, significant event time points are those where the behavioral agent's tolerance for changes exceeds a certain threshold. This threshold is a variable that is influenced by both user profile and event type.

[0042] S5: The planning agent processes the feedback and feeds the processing results back to the planning agent for iterative optimization. The iterative process repeats until the behavioral agent completes the full activity simulation. The planning agent's processing of feedback also includes: if the user chooses to proceed with the current plan, the current activity time and space are maintained, the simulation of the behavioral agent is not stopped, the tolerance of the behavioral agent is reset, and the above simulation steps are continued until the simulation of this activity ends or a new tolerance threshold is reached. Unless otherwise specified by the user, the planning agent will also store the user's choice as user profile information in the local user preference information and update it. If a user suggests modifications, the system will either completely reset the activity timeline or return to the previous timeline node for replanning, based on the provided feedback, and continue simulating until the behavioral agent completes the full activity simulation.

[0043] Understandably, regardless of whether the user chooses to approve or modify the scheme, the system will encrypt and store the interaction data in the local preference library by default. If the user needs to disable this function in the settings, they need to actively disable the update of the user profile. At the same time, to ensure privacy and security, the local preference library is only used for the construction of the overall user profile through federated learning, which will not lead to the leakage of the user's personal privacy.

[0044] Understandably, users should confirm their feedback to the interactive agent within 120 seconds as prompted by the system by clicking the feedback button. If the timeout expires, the system should retrieve the user profile from the local preference library to provide overall feedback, and then continue simulating the activity time. If the user subsequently rejects the feedback, they can interact directly with the interactive agent. After the behavioral agent completes the simulation of the activity time and space point of the current step, the activity time and space should be rolled back to before the feedback from the local preference library, and the activity time and space should be re-planned based on the user's current feedback.

[0045] When the user selects the current option, the threshold of the behavioral agent will be reset.

[0046] S6: The planning agent visualizes the activity's time and space and then hands it over to the interactive agent for display to the user.

[0047] After completing iterative optimization, the planning agent will visualize the activity's spatiotemporal processes, including: First, an activity space map containing the map of the activities is created. The planned activity path, which will be verified by the behavioral agent, is then marked on the map, and spatiotemporal points of the activities are labeled along the route. Each activity spatiotemporal point is labeled with specific time and time period weather information. During interaction, the overall weather environment at the current time is also labeled on the map to help users perceive the overall weather environment of the activity. In addition, the weather map visualized by the planning agent also displays the following: regional route map, road route and type map, and the relationship between road status and recent weather environment. By displaying road information, the travel experience during the activity can be further optimized. For the specific personalized construction form of the behavioral intelligence agent, the modification of the user profile follows the following logic: At inactive points, that is, the spatiotemporal points of movement between activities, the user's traffic preference mode and weather conditions are modeled. In the determined event, the preference weight of the weather environment corresponding to the event is increased significantly, while the preference weight of other non-corresponding weather environments is increased slightly. A coefficient that decays over time is added to the change in amplitude. When the same behavior is no longer repeated, the increase in amplitude decays over time. Whenever the behavior is repeated, the decay parameter is reduced, so that the behavioral intelligence agent can still grasp the user profile well even with a small amount of data. Similarly, the behavioral agent also models the relationship between activity type and meteorological conditions. In a single specific activity event, the meteorological environment preference weight corresponding to the event is increased significantly, while the preference weight of other non-corresponding meteorological environments is increased slightly. A coefficient that decays over time is added to the magnitude change. When the same behavior is no longer repeated, the magnitude increase decays over time. Whenever the behavior is repeated, the decay parameter is reduced, thereby better modeling the relationship between meteorology and activity type. In the planning process of the intelligent agent, the confidence level classification of unofficial information is related to user preferences. Specifically, when the user preference is more aggressive, weather activity plans that do not reach the confidence threshold but are more in line with the user's preferences can also be adopted and planned. However, they need to be highlighted and compared with official information in the activity planning and global visualization. For non-aggressive users, only third-party weather and activity information that matches the official authoritative information is provided.

[0048] The planning agent visualizes the activity space and time, including: dividing the map information of the activity space into multi-level map forms according to the size of the activity space; designing interactive features for specified activity space and time points; and displaying the overall meteorological information of the current map in the form of a meteorological map when interacting with activity space points. The interactive content includes: encountered abnormal meteorological events and meteorological indicators affecting the activity. In summary, the present invention provides a method for constructing and optimizing a meteorological service system based on multi-agent technology. By acquiring user information and additional external meteorological service information, the method achieves modeling and optimization of activity spatiotemporal data through a multi-agent meteorological service system, thereby improving the user experience when querying meteorological services and ensuring the efficiency and reliability of activity spatiotemporal data. This invention achieves planning and optimization of user activities in terms of meteorology through activity spatiotemporal modeling and spatiotemporal node experience. Users can preview and further plan the overall activity through the provided interactive activity spatiotemporal route map, thereby improving the user experience of the meteorological service system.

[0049] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

[0050] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A meteorological service optimization system based on multi-agent technology, characterized in that: include: Interactive agents are used to interact with users and other agents, convey information, clarify needs, and act as intermediaries to enable other agents to effectively understand user needs, while enabling users to quickly understand the weather-based activity plan given by the system as a whole and provide feedback. Meteorological intelligent agents are used to query and organize meteorological information, including professional and accurate real-time meteorological information and spatiotemporal information of events; The planning agent is used to plan the activities provided by the user in the time and space, and optimize the activity time and space by combining the information and feedback provided by the meteorological agent and the behavioral agent. After the maximum number of iterations is reached while meeting the target conditions, the designed activity time and space is output. Behavioral intelligence agents are used to personalize and simulate user behavior logic and provide optimization requests with preferences and specificity, including: simulating user activity planning, using the activity space-time established by the planning intelligence agent as the activity range, experiencing and evaluating the planning in the activity space, and providing nodes to be optimized for iterative optimization.

2. The meteorological service system based on multi-agent technology according to claim 1, characterized in that: The interactive intelligent agent interacts with the user and other intelligent agents, conveying information including: user needs, user-specified modification methods, and the user profile. This information is then exchanged with the meteorological intelligent agent, the planning intelligent agent, the behavioral intelligent agent, and local storage. The feedback from the meteorological and planning intelligent agents regarding the needs is refined by the interactive intelligent agent and fed back to the user. The interactive intelligent agent performs global planning for the system, providing guidance for the overall system.

3. The method for constructing and optimizing a multi-agent-based meteorological service system according to claim 1, characterized in that: The professional and accurate real-time meteorological information and activity spatiotemporal information are obtained through authoritative third-party meteorological agencies. The activity spatiotemporal information is obtained by knowledge base and external Internet search and filtering, or provided by users. The meteorological agent can query authoritative meteorological information provided by credible third-party authoritative meteorological service providers, and can search for meteorological guidance on relevant activities in the activity area through official media, self-media, and social platforms on search engines. At the same time, it can query more detailed meteorological information based on the feedback from the planning agent, providing the planning agent with global information reference and unofficial information.

4. A meteorological service system based on multi-agent technology according to claim 1, characterized in that: The planning agent can combine all received information to optimize and construct the activities proposed by the user, provide a complete solution, and make comprehensive plans based on user experience and personalized preferences.

5. A meteorological service system based on multi-agent technology according to claim 1, characterized in that: The behavioral intelligence agent can simulate the user's behavioral logic and perform personalized construction on the basis of conventional logic, including: simulating conventional user-defined behavioral logic when not receiving a user profile, and reading local storage and modifying the predefined user profile during user use to make the behavioral logic conform to user preferences.

6. A method for constructing and optimizing a multi-agent-based meteorological service system, used in the system described in any one of claims 1 to 5, characterized in that, The method includes the following steps: The interactive intelligent agent interacts with the user to obtain the activities the user needs to perform, and combines the Internet to obtain weather information related to the activities; The meteorological intelligent agent performs queries on user activities and the spatiotemporal meteorological environment; The planning and construction activities of the intelligent agent are planned in time and space, simulating the behavioral logic of users in specific activities, and combining the personal information provided by users to carry out personalized planning and modeling. The behavioral agent simulates activities according to the pre-defined activity plan time and space, provides real-time feedback on the problems existing in the current activity time and space, and the interactive agent receives and organizes the feedback and provides feedback to the user. The planning agent processes the feedback and feeds the results back to the planning agent for iterative optimization. The iterative process repeats until the behavioral agent completes the full activity simulation. The planning agent visualizes the activity's time and space, and then hands it over to the interactive agent for user display.

7. The method for constructing and optimizing a multi-agent-based meteorological service system according to claim 6, characterized in that: The activities that the user needs to perform include: temperature, humidity, wind speed, meteorological conditions and corresponding sensory information during the activity time; the meteorological information related to the activity obtained by combining the Internet includes: the specific content of the activity, the duration of the activity, and meteorological precautions related to the activity; the meteorological agent's query of the user's activities and the spatiotemporal meteorological environment includes: the expected spatiotemporal location of the interactive agent and the activity execution, the dynamic spatiotemporal transition between activities, a spatiotemporal location search performed once every preset time period, and the detection of user status and meteorological environment in subsequent spatiotemporal periods.

8. The method for constructing and optimizing a multi-agent-based meteorological service system according to claim 6, characterized in that: The activity simulation includes: the state at each activity point and movement between different activity points; the behavioral agent performs activity simulation according to the preset activity planning time and space, including: The behavioral agent and the planning agent interact iteratively, and the spatiotemporal activities of the meteorological domain planned by the planning agent are evaluated and optimized through real-time feedback and tolerance accumulation. The evaluation and optimization formula is expressed as follows: in, This refers to the cumulative tolerance level over time. To describe the inappropriateness of the current meteorological environment, t represents time. A negative constant indicates that the cumulative tolerance decreases over time; by summing the tolerance at multiple time points, the total tolerance becomes... If the user exceeds the tolerance threshold, it is considered that the user cannot tolerate the discomfort, the simulation is terminated, and feedback is sent to the planning agent.

9. The method for constructing and optimizing a multi-agent-based meteorological service system according to claim 6, characterized in that: The planning agent's processing of feedback also includes: If the planning agent does not process the feedback or accepts the problem, the current activity spacetime is maintained, and the simulation of the behavioral agent is not stopped until the end of the current activity simulation; if the user proposes modifications, the activity spacetime is completely reset or replanned from the previous spacetime node.

10. The method for constructing and optimizing a multi-agent-based meteorological service system according to claim 6, characterized in that: The planning agent will visualize and output the activity spatiotemporal aspects, including: The planning agent divides the map information of the activity space into multi-level map formats according to the size of the activity space, and specifies interactive design for activity spatiotemporal points. When interacting with activity points, it displays the overall meteorological information of the current map in the form of a meteorological map. The interactive content includes: abnormal meteorological events encountered and meteorological indicators affecting the activity.

Images