Swarm-robot formation optimization based on multiobjective genetic algorithm

On improving the performance in which the swarm-robot grid formation motion are controlled in a complicated circumstance based on virtual force method, it is used the multiobjective genetic algorithm to optimize control parameters. Performance indexes include collision, break the ranks, connectivity, etc. The weight value of the indexes is determined by their importance. Optimization model is established, the multiobjective genetic algorithm based on Pareto sets is used to search the solution of the problem. Simulation results show that this algorithm is effectively capable of obtaining a set of non-dominated solution within a finite evolutionary generation, which overcomes the weakness of handiwork to set control parameters.