Research Multi-path Planning Technology Based On Exclusion And Clustering Niche Genetic Algorithm

Global Path Planning(GPP) is the key technique which decides the Robert's intelligent level.It is a main component of Autonomous Navigation System and studied in this paper. First,the characteristics of path planning problem and genetic algorithm are analyied.Considering the uncertainty of fitness evaluation function in the specific navigation planning problem, and based on the theory of multi-modal function optimization, a conception of multi-path planning is proposed, which can compensate the shortcoming of the fitness evaluation function's uncertainty by providing one best path and some suboptimal paths for the decision-maker.The whole algorithm is divided into two steps: the first one is to obtain population diversity to strengthen the searching capability of the algorithm and realized by exclusion model of niche genetic algorithm.The second one is to obtain the whole population's rapid convergence and realized by the adaptation of clustering algorithm and multiple populations isolation evolving algorithm.Some new viewpoints about exclusion mechanism, clustering algorithm for subpopulation formation and multiple populations evolving parallelized are involved and corresponding coding schemes and genetic operators are given. Optimization is operated at the same time, redundant nodes are removed by simplified operator.Amendment operater assists to weate feasible paths in complex environment and operation improves the speed of path generation.Simulation are performed which show that the designed algorithm can keep good tradeoff between global optimized searching and fast population convergence.Good searching ability and fast convergence can be obtained,which lead to multiple suboptimal paths.The algorithm has considerable value in the path planning research theoretically and practically.