Evolving Cooperative Groups: Preliminary Results
Maria Gordin, Sandip Sen, Narendra Puppala · 1997
Multi-agent systems require coordination of sources with distinct expertise to perform complex tasks effectively. In this paper, we use co-evolutionary approach using genetic algorithms to evolve multiple individuals who can effectively cooperate to solve a common problem. We concurrently run a GA for each individual in the group. In this paper, we experiment with a room painting domain which requires cooperation of two agents. We have used two mechanisms for evaluating an individual in one population: (a) pair it randomly with members from the other population, (b) pair it with members of the other population in a shared memory containing the best pairs found so far. Both the approaches are successful in generating optimal behavior patterns. However, our preliminary results exhibit a slight edge for the shared memory approach. 1 Introduction Our goal is to generate behavior strategies for cooperative agents that have distinct capabilities. A coordinated group effort is necessary to s...