Multi-agent system security convergence control method based on node information privacy protection

Abstract

The invention discloses a multi-agent system security convergence control method based on node information privacy protection, which comprises the following steps of: (1) decomposing an initial state value of an agent, performing state updating for realizing a privacy protection target once, and then constructing a storage vector, after the storage vector of each agent is safely propagated for one time, a screening mechanism with minority obeying majority is used for storing a credible state value vector in each round of iteration, and the state of the agent is updated by using finally received information, so that the state convergence of the agents in all system networks is finally realized; and (2) analyzing conditions for realizing accurate convergence control and privacy protection when the multi-agent system suffers from malicious network attack by applying knowledge of a correlation graph theory and an information theory. Compared with a traditional method, the method breaks through the limitation in the aspects of a multi-agent system which only considers malicious network attacks and does not consider the privacy protection problem and the like, and has a very strong practical prospect in the fields of intelligent industrial application and smart cities.

Description

technical field [0001] The invention belongs to the technical field of distributed multi-agent system coordination control, in particular to a multi-agent system safety convergence control method based on node information privacy protection. Background technique [0002] Multi-agent systems (MASs) are network systems composed of a large number of agents with independent dynamic behaviors. Each agent in the system can perceive the surrounding environment through sensors and communicate with other agents. , to achieve information interaction. The consensus problem is one of the core research directions in the field of distributed cooperative control of multi-agent systems. It means that in the absence of a central node, each node in the system only uses the state information broadcast between neighboring nodes to The dynamic equation of the agent is coupled with the topology of the communication network to form a complex network, and an appropriate distributed control algorit...

AI Technical Summary

Problems solved by technology

Specifically, the security convergence control of a multi-agent system faces two key problems in practical applications: (1) The traditional convergence control method directly transmits the state value of the agent itself through an open channel, which can easily lead to the privacy leakage of the sensitive information of the node itself; (2) The communication links between nodes may be subject to malicious network attacks, such as denial-of-service (DoS) attacks, deception attacks, etc. This type of network attack is a common type of damage to the system means of overall stability

Images