CAJ | 학술논문

应用蚁群生成树算法搜索了有34个节点的连接图的生成树，并采用正交设计法和均匀设计法进行了参数优化配置方法研究。结果表明：对于参数较多的蚁群算法，应用正交设计法和均匀设计法进行参数优化配置是一种可行且有效的途径，可有效提高蚁群算法的收敛速度，在求解精度上也有一定优势；充分发挥人类智能与仿生物智能的各自优势是克服单纯靠智能优化方法随机搜索缺点的关键；当蚂蚁数目为100、信息素相对重要性因素为0．3、信息素衰减系数为3．6、信息素挥发系数为0．4、信息素增加强度系数为14时，蚁群生成树算法效果最佳。
응용의군생성수산법수색료유34개절점적련접도적생성수，병채용정교설계법화균균설계법진행료삼수우화배치방법연구。결과표명：대우삼수교다적의군산법，응용정교설계법화균균설계법진행삼수우화배치시일충가행차유효적도경，가유효제고의군산법적수렴속도，재구해정도상야유일정우세；충분발휘인류지능여방생물지능적각자우세시극복단순고지능우화방법수궤수색결점적관건；당마의수목위100、신식소상대중요성인소위0．3、신식소쇠감계수위3．6、신식소휘발계수위0．4、신식소증가강도계수위14시，의군생성수산법효과최가。
The ant colony spanning tree algorithm were used to find the Spanning Tree of the graph which had 34 nodes and the orthogonal design method and the uniform design method were used to optimize its parameters. The results show that for the Ant Colony Algorithm with many parameters,it is a useful and effective way to determine the parameters combination by applying the orthogonal design method and the uniform design method,and it can effectively improve the algorithm convergence and has some advantages in computational accuracy. The key to overcome the shortcomings of only random search relying on intelligent optimization algorithms is to take advantage of the human intelli-gence and bionic intelligence. When the ant numbers is 100,the relative importance of factors of pheromone Beta is 0. 3,the decay coeffi-cient of pheromone Alpha is 3. 6,the evaporation coefficient of pheromone Rho is 0. 4 and the strength coefficient of pheromone Qt is 14,the efficiency of the ant colony spanning tree algorithms is the best.