A model of agent persuasion based on genetic algorithms: Design considerations

Reaching agreements is an important problem in multi-agent systems (MAS), which requires different types of dialogues between agents. Persuasion is one of them and it is predominantly based in first order logic. However, this type of dialogue is not an adequate approach for solving complex problems - those that can't be solved in polynomial time. This paper presents a persuasion model for MAS which consider the fitness function of a genetic algorithm (GA) for reaching agreements in problem solving. The resulting persuasion model provides each member of MAS with a fitness function capable of guiding agents in the reach of agreements. The evaluation we did enabled us to state that the model could have a positive impact for solving complex problems with MAS. In contrast with previous studies, our results highlight the importance of using a GA persuasion model for solving complex problems and we depict the implications of our findings in its design considerations.