Algorithm for Reducing the Intensity of Conflicts in Hybrid Intelligent Multi-Agent Systems

The paper is devoted to the developing of the algorithm for reducing the intensity and, if possible, resolving conflicts that arise in hybrid intelligent multi-agent systems. This algorithm is a necessary component of the conflict management method in such systems. Conflict management in hybrid intelligent multi-agent systems involves identifying a decision-making situation, stimulating, if necessary, constructive forms of conflict and then resolving them using the proposed algorithm, as well as preventing its destructive forms. Given the heterogeneous nature of the system in terms of knowledge areas modeled and problem-solving methods used, conflict stimulation allows the system to consider the problems posed from different points of view. The use of the proposed algorithm for reducing the intensity and resolving conflicts after execution of the conflict stimulation methods and generation of a large number of problem solutions allows agents to automatically bring their positions closer and develop a single collective solution, thus saving the user from having to compare alternatives manually.