一种混合智能排样优化算法的应用研究

A Study on Packing Problem Based on Hybrid Intelligent Optimization Algorithm

基于智能优化方法的混合机理,将并行进化机制引入遗传算法和粒子群算法,提出一种混合智能优化排样方法(HGPA)。该算法依据个体适应度值的大小和相似性对整个种群进行合理划分,在每次迭代中,个体适应度值较好的子种群利用遗传算法进化,个体适应度值较差的子种群则利用粒子群算法处理,实现优化方法的优势互补和信息增值。同时通过设置多样性度量标准来控制种群特征信息和搜索空间。在求解不规则件排样问题的算例表明:该算法可平衡控制个体种群进化中的局部寻优和全局搜索,为智能优化的混合机理研究提供了一个新的思路。

Based on hybrid mechanism and strategies,a hybrid intelligent optimization algorithm（ HGPA） is proposed by introducing the parallel evolution mechanism into genetic algorithm（ GA） and particle swarm optimization（ PSO）. In this algorithm the individuals are divided into two small groups according to their fitness values and similarity. In the iteration,the sub-group of the top fitness values is evolved by GA,and the other subgroup is evolved by the PSO algorithm to achieve complementary advantages and value-added information. The evaluating standard of diversity is set to control the characteristics information of population and the searching space.The irregular parts packing experiment show that the HGPA can evenly control the evolution of individual local optimization and global search,and provide a new method of hybrid intelligent optimization.