Intelligent Multiagent Coordination Based On Neuro-Fuzzy Models With Hierarchical Reinforcement Learning

Abstract

Abstract-This paper presents the research and development of a hybrid neuro-fuzzy model for the hierarchical coordination of multiple intelligent agents. The main objective of the models is to have multiple agents interact intelligently with each other in complex systems. We developed two new models of coordination for intelligent neuro-fuzzy multiagent systems that use MultiAgent Reinforcement Learning Hierarchical Neuro-Fuzzy with a market-driven coordination mechanism (MA-RL-HNFP-MD) and a MultiAgent Reinforcement Learning Hierarchical Neuro-Fuzzy with graph coordination (MA-RL-HNFP-GC). After setting options and using the MA-RL-HNFP_MA family of models, the coordination systems were tested in two case studies involving the implementation of a benchmark game of predator-prey. The tests showed that the new system has the ability to coordinate actions between agents with a convergence rate nearly 30% greater than that of the original version.