改进的基于蜜蜂进化型遗传算法和蚁群系统混合的元件贴装优化

The Component Mounting Optimization Based on Improved Bee Evolutionary Genetic Algorithm and Ant Colony System

针对PCB板的表面贴装技术(Surface Mount Technology,SMT)优化问题,提出一种基于蜜蜂进化型遗传算法和蚁群系统的混合智能算法(the Hybrid Intelligent Algorithm based on Bee Evolutionary Genetic Algorithm and Ant Colony System,BAHA).该算法的关键有4点:①通过两个种群的融合实现信息共享,提高算法的收敛速度;②采用改进的OX的交叉算子,合理保留优秀个体基因的排列顺序;③加入局部搜索算子,在当代最优解附近进行更加精细的搜索;④信息素重置防止陷入局部最优解.用TSP30问题、eil51问题与相关文献进行对比测试,仿真结果表明BAHA收敛速度快,寻优能力强.通过对5种不同PCB板的元件贴装顺序进行优化计算,结果表明,BAHA能有效的提高贴装效率.

To optimize the Surface Mount Technology （SMT） of PCB card, a hybrid intelligent algorithm based on bee evolutionary genetic algorithm and ant colony system （BAHA） is proposed. The key of the hybrid intelligent algorithm lies in improving the convergence speed by the combination between the two population; the improved OX crossover operator retained the sequence of the good genes; introducing a Local search operator, which has more elaborate search ability in the neighborhood of the iteration-best~ pheromone resetting is used to jump from local optimal solution. The test results on TSP 30 problem, ell51 problem and the component mounting sequence problem of 5 PCB cards show that BAHA has good global search ability and fast convergence rate, and promote the mounting efficiency effectively.