摘要
基于传统遗传算法在移动机器人路径规划中应用的不足,对遗传算法进行了一定的改进。在初始化种群中采用闵科夫斯基和原理扩展障碍物,选择真正可行的区域,在可行区域中去初始化种群,这样提高了进化的速度;在选择算子中引入了相似性的概念,扩大父代的种类,避免快速进入局部最优解;在交叉算子中采用了动态确定变异概率,这样可以提高个体的质量;通过仿真证明了改进的遗传算法能够更快的收敛到全局最优解,方法是正确有效的。
It makes some improvements on the Genetic Algorithm based on the deficiencies of tra-ditional Genetic Algorithm used in the robot path planning,which is that the use of principle of Minkowshi to expand the obstacles,then to select real feasible region to initial population,which can improve the speed of evolution;the introduction of the concept of similarity in the selection operator,then to expand the type of parent,which can avoid local optimal solution quickly;and the application of the dynamic mutation probability in the crossover operator,which improves the quality of individual;the method is correct and ef-fective through the simulation,which proves the improved Genetic Algorithm can converge more quickly to the global optimal solution.
出处
《机械设计与制造》
北大核心
2010年第7期147-149,共3页
Machinery Design & Manufacture
