建筑运维智能体策略更新方法及存储介质

By introducing the Shapley value credit allocation mechanism into the intelligent building operation and maintenance system, the problem of unfair credit allocation in multi-agent systems is solved, and optimization and dynamic adaptation of cross-system collaboration are achieved, thereby improving operation and maintenance efficiency and fairness.

CN121920409BActive Publication Date: 2026-07-17SHANGHAI CONSTRUCTION FOURTH CONSTRUCTION GROUP CO LTD

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHANGHAI CONSTRUCTION FOURTH CONSTRUCTION GROUP CO LTD
Filing Date
2025-12-23
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

In intelligent building operation and maintenance management systems, existing technologies struggle to effectively assess the cross-system collaboration value among multiple intelligent agent systems, leading to unfair credit allocation and a lack of ability to model dynamic collaborative relationships, making it difficult to adapt to changes in the operating environment caused by building usage patterns and equipment aging.

Method used

A credit allocation mechanism based on Shapley values ​​is adopted. By training the joint action value function, the Shapley credit value of each agent is calculated. Combined with the operation and maintenance collaboration affinity matrix and adaptive sampling strategy, the action strategies of each agent are optimized.

Benefits of technology

It enables fair evaluation of the contributions of each intelligent agent, enhances cross-system collaboration and optimization capabilities, dynamically adapts to building operation needs, and improves overall operation and maintenance efficiency and fairness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121920409B_ABST
    Figure CN121920409B_ABST
Patent Text Reader

Abstract

本发明提供了一种建筑运维智能体策略更新方法及存储介质,本发明充分考虑暖通空调、安防监控、设备维护、环境控制等子系统间的协作价值,准确量化各专业智能体对建筑整体运营效果的贡献;将合作博弈论中的沙普利值理论引入建筑运维CTDE框架,为各运维智能体提供理论公平的信用评估,解决传统方法在多目标权衡中的公平性问题;针对大型建筑物可能包含数十个专业智能体的情况,提供高效的沙普利值近似计算方法,使其在实际建筑运维系统中具备可行性;通过精确的信用分配指导各运维智能体的策略优化,实现建筑物能耗、舒适度、安全性、维护成本的综合最优化。
Need to check novelty before this filing date? Find Prior Art