AI multi-agent-based old cell reconstruction decision-making method and system

Abstract

The invention relates to the technical field of artificial intelligence, in particular to an old community reconstruction decision-making method and system based on AI multiple agents, and the method comprises the steps: S1, collecting opinions of all parties through a data collection module; s2, constructing an AI agent representing opinions of all parties through an agent interaction module; s3, a decision optimization module adopts a polling speaking rule based on an attention mechanism to drive AI agents of all parties to perform multiple rounds of interactive speaking; s4, a decision optimization module promotes AI agents of all parties to continuously adjust the initial reconstruction scheme in multiple rounds of interactive speaking through a dynamic game algorithm, and a final reconstruction scheme is formed. Mass data can be rapidly processed, appeals of all parties can be integrated in a short time, all parties can communicate with one another, and the method and the device have the advantages of being high in practicability and easy to popularize. Therefore, the feasibility of the reconstruction scheme is improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and particularly to a decision-making method and system for the renovation of old communities based on AI multi-agents. Background Art

[0002] In the complex social engineering field of the renovation of old communities, researchers' timely and accurate collection, collation, and analysis of public opinions play a crucial cornerstone role in constructing a practical social collaborative governance plan. However, the traditional working mode has many drawbacks in this key link.

[0003] For a long time, the information collection and collation links have highly relied on manual operations. Researchers need to spend a lot of time and energy shuttling through various old communities, collecting residents' demands and feedback on aspects such as water supply, electricity, property services, infrastructure maintenance, environmental problems, safety hazards, housing quality, and community management through methods such as questionnaires, on-site visits, and telephone interviews. However, manual operations are not only inefficient, prone to situations such as misrecording and missing recording, resulting in a high error rate, but also, during the collation process, affected by subjective factors such as the individual knowledge structure and cognitive biases of researchers, it is difficult to conduct scientific and accurate quantitative analysis on the complex data, greatly reducing the quality of the initial data.

[0004] Even if the information collection and collation are difficultly completed, the subsequent analysis link also faces severe challenges. Relying on the limited experience cases accumulated in the past, researchers start to analyze the interest demands of all parties such as local governments, sub-district offices, property companies, neighborhood committees, and community owners, and accordingly try to propose solutions. However, the number of past experience cases is limited and cannot cover the ever-changing actual situations in the renovation of old communities, making it difficult to provide comprehensive and targeted references. At the same time, researchers mostly work behind closed doors based on second-hand materials in the office, lacking face-to-face and in-depth communication with the real interest subjects and failing to let all parties truly participate in the real game process. This leads to the proposed solutions often being out of touch with reality, having great limitations, unable to fully balance the interests of all parties and truly meet the needs of all parties, ultimately making it difficult for the renovation of old communities to achieve the ideal collaborative effect and the satisfaction of residents to be substantially improved. Therefore, the present application proposes a decision-making method and system for the renovation of old communities based on AI multi-agents. Summary of the Invention

[0005] To solve the above problems, the present invention provides a decision-making method and system for the renovation of old communities based on AI multi-agents. Through the present application, a large amount of data can be quickly processed, the demands of all parties can be integrated within a short time, and all parties can communicate with each other, taking into account the interests of all parties to improve the feasibility of the renovation plan.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A decision-making method for the renovation of old residential areas based on AI multi-agent, comprising:

[0008] S1: Collect opinions from all parties through a data collection module;

[0009] S2: Construct AI agents representing the opinions of all parties through an agent interaction module;

[0010] S3: The decision optimization module adopts a speaking rule based on the attention mechanism to drive the AI agents of all parties to conduct multi-round interactive speeches;

[0011] S4: The decision optimization module uses a dynamic game algorithm to prompt the AI agents of all parties to continuously adjust the initial renovation plan during multi-round interactive speeches to form a final renovation plan.

[0012] Preferably, after S2, it further includes:

[0013] S2.1: Based on the big data collected by the data collection module, the decision optimization module uses clustering analysis and association rule mining algorithms to initially identify the key problem areas of old residential areas;

[0014] S2.2: According to the key problem areas, the decision optimization module automatically retrieves the expert knowledge base and Internet messages, extracts relevant case experiences and technical parameter information, and generates an initial renovation plan.

[0015] Preferably, after S2.2, it further includes: During each round of interactive speech of the AI agents of all parties, the speaking time is dynamically allocated according to the key problem areas of the current renovation plan.

[0016] Preferably, S3 further includes:

[0017] Introduce a dynamic reputation evaluation mechanism to record the integrity behaviors of the AI agents of all parties during multi-round interactive speeches, and update the reputation scores of the AI agents of all parties in real time;

[0018] Allocate decision-making power according to the reputation scores of the AI agents of all parties.

[0019] Preferably, before S4, it further includes:

[0020] After several rounds of interaction, the decision optimization module comprehensively evaluates the current renovation plan;

[0021] If the evaluation result does not reach the preset threshold, trigger plan iteration, return to S3, until a final renovation plan is generated.

[0022] Preferably, the criteria for the decision optimization module to comprehensively evaluate the current renovation plan include the expected improvement in residents' satisfaction, the input-output ratio of government funds, the profit prediction of property management companies, and the achievement degree of the street office's political achievements.

[0023] An old community renovation decision-making system based on AI multi-agent, comprising:

[0024] A data collection module, which is used to obtain the data involved in the renovation of old communities, and analyze, process and store the data;

[0025] An agent interaction module, which is used to create AI agents representing the opinions of all parties;

[0026] A decision optimization module, which is used to drive the AI agents of all parties to conduct multiple rounds of interaction, and generate a final renovation plan according to the results of multiple rounds of interaction.

[0027] Preferably, the decision optimization module is communicatively connected to an expert knowledge base, and the expert knowledge base includes successful and failed cases of past old community renovations, industry standards and technical specifications.

[0028] Preferably, the AI agent includes an interest appeal model, a decision-making model, a communication model, an emotion analysis model and a situation awareness model.

[0029] The beneficial effects of the present invention are as follows:

[0030] Compared with the existing manual-dependent method, the data collection module can quickly process a large amount of data, integrate the demands of all parties in a short time, communicate with multiple parties through the establishment of AI agents of all parties, and through the dynamic game algorithm, prompt the AI agents of all parties to continuously adjust their strategies in multiple rounds of interaction, taking into account the interests of all parties, truly meeting the needs of each subject, and finally forming a final renovation plan to improve the feasibility of the renovation plan. Description of the Drawings

[0031] Figure 1 It is a flowchart of the old community renovation decision-making method based on AI multi-agent in a specific embodiment of the present invention;

[0032] Figure 2 It is a flowchart of generating an initial renovation plan in a specific embodiment of the present invention;

[0033] Figure 3 It is a diagram of the interaction speech rules of AI agents of all parties in a specific embodiment of the present invention;

[0034] Figure 4 It is a diagram of a successful renovation plan in a specific embodiment of the present invention;

[0035] Figure 5This is a scoring diagram of the evaluation mechanism in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0036] Example 1

[0037] See also Figures 1 - 5 As shown, the present invention relates to a decision-making method and system for the renovation of old residential areas based on AI multi-agents.

[0038] An AI agent is a software or hardware system that can perceive the environment, make decisions based on the perceived information, and perform actions to achieve specific goals. It can be fully autonomous or work under the supervision of human operators. The agent uses built-in algorithms to parse data, make plans, and take actions to complete the predetermined tasks.

[0039] Through the data collection module, we gather opinions from all parties, break through the limitation of relying solely on residents' feedback, build a comprehensive opinion collection network, and widely incorporate the voices of all parties.

[0040] In this embodiment, residents can describe the existing problems in the community in detail in a text box on the convenient Internet, such as "some roads in the community are in disrepair, and the bumps and unevenness affect travel" and "there are too few leisure benches, making it inconvenient for residents to rest." The platform uses intelligent guidance functions to automatically classify residents' feedback based on keywords. For example, if the word "road" is mentioned, it will be automatically classified into the infrastructure category to facilitate subsequent integrated analysis.

[0041] In this embodiment, the agent interaction module builds five core AI agents based on cutting-edge deep learning and the latest multi-agent collaborative learning algorithm, simulating the local government, street office, property company, neighborhood committee, and community owners respectively;

[0042] In some other embodiments, the number and role positioning of AI agents may be arbitrary.

[0043] Each AI agent not only has a built-in interest model, decision-making strategy model, and communication model for the corresponding role, but also incorporates a sentiment analysis module and a situational awareness module.

[0044] For example, the interest demand model of the local government AI agent considers factors such as financial budget constraints, people's livelihood improvement goals, and urban image enhancement items. The decision-making strategy model formulates strategies such as fund allocation and project priority ranking based on past similar community renovation cases and policy orientations. Its sentiment analysis module can capture the emotional tendencies of the public towards the renovation policy, and the situation awareness module can dynamically adjust the strategy focus according to factors such as urban development planning and regional positioning. The community owner AI agent constructs a preference model based on the analysis of big data of residents' feedback, focusing on the solution degree of issues closely related to daily life such as water supply, property management, and environment. The communication model simulates the expression habits and key points of demands of the owners during community discussions and complaints. The sentiment analysis module can judge the satisfaction fluctuations of the owners with the renovation plan, and the situation awareness module can adjust the demand weights in real time in combination with seasonal changes (such as high demand for power supply stability in summer) and community activity arrangements (such as increased demand for public space during large-scale events).

[0045] The above AI agents are instantiated in the AI server, with independent running threads and memory spaces, and are isolated and deployed using containerization technology to ensure operation stability and the rationality of resource allocation, and can carry out interactive activities in parallel.

[0046] The decision optimization module adopts a polling speaking rule based on the attention mechanism, and the interaction rule is as Figure 3 shown. In each round of interaction, according to the key areas of the current renovation plan discussion, the speaking time is dynamically allocated to ensure that key information is fully communicated.

[0047] For example, when discussing the greening renovation of old communities, the neighborhood committee AI agent and the community owner AI agent, due to their direct connection with the residents' living experience, can obtain relatively long speaking times to fully elaborate on the residents' expectations for greening varieties and layouts.

[0048] The decision optimization module uses a dynamic game algorithm to prompt each party's AI agent to continuously adjust the initial renovation plan in multiple rounds of interaction to form the final renovation plan.

[0049] Furthermore, after several rounds of interaction, the decision optimization module comprehensively evaluates the current renovation plan. In this embodiment, the evaluation criteria include:

[0050] The expected improvement in residents' satisfaction, through the analysis of the feedback from the community owner AI agent;

[0051] The input-output ratio of government funds, combined with the budget of the local government AI agent and the expected social benefits;

[0052] The profit prediction of the property company, based on the cost-benefit model of the property AI agent;

[0053] The performance achievement of the sub-district office is referenced by the improvement goals for community management set by the sub-district office's AI agent;

[0054] The achievement status of the neighborhood committee's service goals is based on the neighborhood committee's AI agent's tracking of the protection of owners' rights and interests;

[0055] If the evaluation result does not reach the preset threshold, the solution iteration is triggered, and it returns to the intelligent agent interaction module for further optimization until the final renovation solution is generated.

[0056] Refer to Figure 4 As shown, this is the successful renovation solution of this application.

[0057] Embodiment 2

[0058] Preferably, the built-in interface of the decision optimization module is connected to the expert knowledge base. At the initial stage of interaction startup, the decision optimization module, based on the big data aggregated by the data collection module, uses algorithms such as clustering analysis and association rule mining to initially identify the key problem areas of old communities (for example, if a certain community frequently has water outages and there are concentrated negative reviews on property services, then water supply and property management are listed as key renovation areas), and combines with the general renovation templates in the expert knowledge base to generate an initial renovation solution. This initial solution covers the list of infrastructure renovation projects, preliminary allocation of funds, evaluation of expected effects, etc., and serves as the starting point for the interaction of AI agents, promoting all parties to discuss and optimize around the initial solution.

[0059] Among them, the expert knowledge base contains rich knowledge resources such as successful and failed cases of past old community renovations, industry standards, and technical specifications. In the dynamic game process, when the AI agent encounters complex decision-making problems (such as the applicability of new building materials in the reinforcement of old building structures, the cost performance evaluation of different elevator brands), the decision optimization layer automatically retrieves the expert knowledge base and Internet messages, extracts relevant case experiences, technical parameters and other information, provides decision-making references for the AI agent, and enhances the scientificity and reliability of the solution. For example, when encountering the problem of roof leakage, the knowledge base can provide comparison data such as the implementation cost, service life, and maintenance difficulty of various waterproof treatment solutions to assist all parties' AI agents in negotiating to determine the best solution.

[0060] Preferably, during the interaction of the AI agent, a dynamic reputation evaluation mechanism is introduced, where the evaluation score is set between 0-10 points. Refer to Figure 5 As shown in the example, through the evaluation mechanism, record the integrity behaviors of the AI agent during the interaction process (such as the authenticity of data provided, the coherence of following interaction rules), and update the reputation points in real time. The reputation points are directly related to key decision-making powers such as its right to speak and the priority of resource acquisition. For AI agents with good reputations, give preferential treatment in resource allocation (such as computing resources, data access permissions) to encourage honest interaction.

[0061] For example, the AI ​​agent of the street office, which consistently and truthfully provides data on the aging of community facilities, has an advantage in obtaining resources for deploying additional monitoring equipment, and promotes the transformation plan to evolve in the optimal direction that takes into account the interests of all parties.

[0062] Example 3

[0063] Preferably, in some other embodiments, the present application may also be provided with an execution monitoring module, which directly connects the data acquisition module with the execution monitoring module, bypasses the intelligent agent interaction module and the decision optimization module, and constructs a simplified linear process. After data collection, the data related to emergency facility maintenance is directly passed to the execution team based on a preset simple priority model without going through complex intelligent agent games and decision optimization.

[0064] Preferably, in some other embodiments, the present application includes a simple fixed speaking order and public information exchange mode, for example, speakers are made in a fixed role order (such as owner-property-government), each speaking time is fixed, information exchange is not encrypted, and is completely transparent and open, without considering the impact of credibility factors on decision-making weights.

[0065] As a preference, in some other embodiments, the technical route of real-time integration of dynamic game algorithm and expert knowledge base is abandoned, and a static planning model combined with a regularly updated case library is adopted. That is, several fixed transformation plan templates are formulated in advance based on experience or general standards. When encountering a specific community transformation, the plan with the highest matching degree is selected from the template, and the regularly updated case library is fine-tuned, and real-time intelligent agent interactive game and knowledge retrieval-driven optimization are no longer performed.

[0066] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A decision-making method for the renovation of old residential areas based on AI multi-agents, characterized in that: include: S1: Collect opinions from all parties through the data collection module; S2: Build an AI agent that represents the opinions of all parties through the agent interaction module; S3: The decision optimization module uses speech rules based on the attention mechanism to drive the AI ​​agents of all parties to conduct multiple rounds of interactive speeches; S4: The decision optimization module uses a dynamic game algorithm to encourage the AI ​​agents of all parties to continuously adjust the initial transformation plan in multiple rounds of interactive speeches to form a final transformation plan.

2. The old residential area renovation decision-making method based on AI multi-agent according to claim 1 is characterized in that: After S2, it also includes: S2.1: The decision optimization module uses cluster analysis and association rule mining algorithms based on the big data collected by the data collection module to preliminarily identify the key problem areas of old communities; S2.2: Based on the key problem areas, the decision optimization module automatically searches the expert knowledge base and Internet messages, extracts relevant case experience and technical parameter information, and generates an initial transformation plan.

3. The old residential area renovation decision-making method based on AI multi-agent according to claim 2 is characterized in that: After S2.2, it also includes: in each round of interactive speech, the AI ​​agents of each party dynamically allocate speaking time according to the key problem areas of the current transformation plan.

4. The old residential area renovation decision-making method based on AI multi-agent according to claim 1 is characterized in that: The S3 further includes: Introduce a dynamic reputation evaluation mechanism to record the integrity of each party's AI agent during multiple rounds of interactive speeches, and update the reputation points of each party's AI agent in real time; Decision-making power is allocated based on the reputation points of each party’s AI agents.

5. The old residential area reconstruction decision-making method based on AI multi-agent according to claim 1 is characterized in that: Before S4, it also includes: After several rounds of interaction, the decision optimization module conducts a comprehensive evaluation of the current transformation plan; If the evaluation result does not reach the preset threshold, the solution iteration is triggered and returns to S3 until the final transformation solution is generated.

6. The old residential area reconstruction decision-making method based on AI multi-agent according to claim 5 is characterized in that: The decision optimization module comprehensively evaluates the current renovation plan based on the following criteria: expected improvement in residents' satisfaction, government funding input-output ratio, property company profit forecast, and the achievement of the subdistrict office's political performance.

7. An old residential area renovation decision-making system based on AI multi-agent, characterized in that: include: A data acquisition module, which is used to obtain data related to the renovation of old residential areas, and analyze, process and store the data; An agent interaction module, wherein the agent interaction module is used to create an AI agent representing the opinions of all parties; The decision optimization module is used to drive the AI ​​agents of all parties to conduct multiple rounds of interactions and generate a final transformation plan based on the results of the multiple rounds of interactions.

8. The old residential area reconstruction decision-making system based on AI multi-agent according to claim 6 is characterized in that: The decision optimization module is in communication connection with an expert knowledge base, which includes past success and failure cases of old residential area renovation, industry standards and technical specifications.

9. The old residential area reconstruction decision-making system based on AI multi-agent according to claim 6 is characterized in that: The AI ​​agent includes an interest appeal model, a decision-making model, a communication model, a sentiment analysis model and a situational awareness model.

Images