Cluster intelligence system protection method based on specified time distribution aggregation optimization
By designing a local cost function and a time-specified distributed aggregation optimization algorithm, the problem of collaborative protection among agents in a clustered intelligent system is solved, enabling agents to reach the global optimal position within a specified time and improving task execution efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- RES & DEV INST OF NORTHWESTERN POLYTECHNICAL UNIV IN SHENZHEN
- Filing Date
- 2026-06-11
- Publication Date
- 2026-07-17
AI Technical Summary
In the existing distributed optimization framework, the private cost function of each agent in the cluster intelligence system depends only on its own decision variables, which makes it difficult to solve the problem of collaborative protection among agents and makes it impossible to reach the globally optimal protection position at a specified time.
By designing a local cost function, the collaborative protection problem of a cluster intelligent system is transformed into a distributed aggregation optimization problem based on an undirected connected graph. Combined with a time-specified distributed aggregation optimization algorithm, an agent model is constructed using a communication topology graph and a Laplace matrix, enabling the agent to reach the global optimal position within a specified time.
The intelligent agent can find the optimal solution to the distributed aggregation optimization problem within a specified time, accurately complete the collaborative protection task, and improve the efficiency of the swarm intelligent system in tasks such as marine environmental monitoring and unmanned vehicle ground patrol.
Smart Images
Figure CN122414725A_ABST