基于自适应退火遗传算法的船舶管路布局优化方法

采用自适应遗传算法来确定标准遗传算法的杂交率和变异率,尤其对变异率的调整,使其不但能根据个体适应值的大小进行自适应修正,而且能随进化状态的改变而改变,从而增强了算法摆脱局部最优解的能力.同时引入模拟退火思想,通过对标准遗传算法接受算子的退火处理,使其在搜索过程中除了接受优化解以外还以Metropolis准则接受恶化解,提高了种群的多样性,有效地增强了全局寻优能力.通过对适应值函数的退火拉伸,调整了进化前后期的适应值差异,从而加速了寻优过程.最终以形成的自适应模拟退火遗传算法进行船舶管路的三维布局优化,仿真实验表明,该算法不但加快了寻优速度,而且与标准遗传算法相比全局收敛率提高了近30%.

Approach of ship pipe paths routing optimization based on adaptive annealing genetic algorithm

Adaptive genetic algorithm(AGA) is adopted to determine crossover probability and mutation probability of simple genetic algorithm.In particular,by means of analyzing the effects of mutation probability on global convergence performance of genetic algorithm,the improved method changes mutation probability automatically in accordance with not only fitness value of each individual but also different evolution states.In the mean time,simulated annealing algorithm(SAA) is introduced to modify the acceptance operator and fitness function value.The new acceptance operator accepts not only better individuals but also worse ones at Metropolis probability.This acceptance operation enhances the diversity of population and the capability of search for the global optimum.The stretched fitness function value improves the distribution of fitness values during the prophase and anaphase of evolution,which accelerates the evolution process of algorithm.An adaptive simulated annealing genetic algorithm(ASAGA) is formed and used to optimize the ship pipe paths in 3D modeling space.The simulation results demonstrate that the approach improves the performance in searching speed and increases the global astringency by 30% compared with simple genetic algorithm(SGA).