微型电脑应用

以往求解多旅行商问题的研究仅局限于以各旅行商路程总和最小为优化标准的传统遗传算法,而没有考虑他们的速度和所花时间。在MPI并行环境下,用C++语言实现了粗粒度模型的并行遗传算法。结合并行遗传算法的特点,提出了解决多旅行商问题的策略以及给出相应的算法过程,并进行了有效验证。通过研究结果表明,与传统遗传算法相比,并行遗传算法提高了运算速度,降低了平均开销时间并且最小总路径值更理想。

Previous reseaches on the problem of multiple traveling salesmen were only limited to that employed total-path-shortest as the evaluating rulefor the traditional genetic algorithm, without consider their speed and time spent. Under MPI, parallel genetic algorithms based on coarse-grained modelwas realized in this article,which C++ language was used. Combining the characteristics of parallel genetic algorithm,a strategy of multiple travelingsalesman problem was proposed,the corresponding process of algorithm were also given, and an effective verification was conducted. The results showthat parallel genetic algorithm improves the computing speed,reduces the average cost of time and the total-path-shortest is better compared with thetraditional genetic algorithm.