基于混合群集智能算法的并行公差优化设计

Concurrent tolerance optimization design based on hybrid swarm intelligence algorithm

通过对并行公差优化设计的分析,将其视为一种混合变量组合优化问题。首先给出了并行公差优化设计的数学模型,然后将其映射为一类特殊的旅行商问题——顺序多路旅行商问题,从而降低了问题的求解难度。利用蚁群优化算法和粒子群优化算法,分别在求解离散和连续变量优化时的优势,提出了一种求解并行公差优化设计问题的混合群集智能算法。通过一个计算实例,将混合群集智能算法分别与遗传算法和模拟退火算法进行了比较,结果表明,前者具有更强的搜索能力和较高的效率。同时,混合群集智能算法也为求解一般意义的混合变量优化问题提供了借鉴和参考。

According to the analysis of the characteristics of concurrent tolerance optimization design, it could be regarded as the hybrid variable combinational optimization problem. Firstly, the mathematical model of concurrent optimization was presented. Then it was mapped to a kind of special Traveling Salesman Problem （TSP）-the Sequence Multi-way Traveling Salesman Problem （SMTSP）, which could reduce the difficulty of solving problem. Borrowing the advantages of Ant Colony Optimization （ACO） in discrete problems and Particle Swarm Optimization （PSO） in continuous problems, a Hybrid Swarm Intelligence Algorithm （HSIA） was proposed, which was applied to concurrent tolerance optimization design and searching satisfied results. Result of a simple computational example showed that the HSIA was more efficient with stronger searching ability than Genetic Algorithm （GA） and Simulation Annealing （SA） algorithm. Furthermore, it also provided a new idea to solve the general hybrid variable optimization problem.