Evolutionary Multi-Agent Model With Intent Exchange Solving Multi Objective Optimization Problem

In this paper, an evolutionary Multi-Agent model based with intent exchange is presented to solve multi objective optimization problem. There is no center control in this model, every agent is independent. Each agent has different "intent", which includes preferences for different objects, vectors of movement and evaluations for these vectors. Agents learn from feedback of searches. Agents also exchange intent with neighborhoods to get better performance. Agents will evolve base on search history. Finally, several simple multi objective optimization problems are considered to evaluate this model. The experimental results prove the valid of this model.