Select Reliable Strategy in Multi-Agent Systems Using Fuzzy Logic-Based Fusion

Abstract

In this paper we implemented a fully knowledge base approach for determine the attack/defense strategies of a team of soccer robots as a multi agent system which are playing two cooperative and non-cooperative games in a battle field simultaneously. First two teams play in non-cooperative game and find their equilibrium point and payoff then all players in a team tries reaching to this position. As there are several decision about team strategy in non-cooperative game and player strategy in cooperative game, our novel approach to use Fuzzy Integral Operators (F10) to fuse sensory information enables designer to softly change team strategy and play the most likelihood game to a real game. Selecting the shape of FIO affect directly on the team strategy and players' position and be reviewed in this paper. This approach will be imported in both, cooperative and non-cooperative game when we face to a multi- hypothesis structure. Results justified successfully the improved performance in team action.