Computer-based strategy development in multi-agent systems using genetic algorithms

The task of developing a strategy that generates highly adaptable and collaborative agents and produces optimum results can be an exhaustive process. Changes on the system environment conditions can make this task even more difficult. Using genetic algorithms to solve such tasks can be a reasonably fast and lightweight solution and offer a few advantages over traditional approaches that use Reinforcement Learning.; This work applied a genetic algorithm to a multi-agent system with the goal of examining techniques of strategy development applied to particular problems or goals common to a set of agents. The approach was experimented and the results showed that a genetic algorithm could generate an optimum result and overcome various optimization issues faced by other approaches.