智能多目标网络设计软件系统的研制

Research of intelligent multi-objective network optimum design software

讨论了用遗传算法求解多目标的网络优化问题,以获得最优解或近优解。在算法设计中,使用了非主宰分类遗传算法进行求解。在编码设计时,使用路径所经过的节点序列作为染色体编码,编码长度为节点的总个数,未经过的节点在染色体序列中的相应位置设为零。在执行算法操作时,采用联赛选择、部分匹配交叉、交换变异和优值保留相结合的方法来进化取优。软件设计时,使用VC++程序设计语言,实现了相关信息的图文显示,一些算法参数可以修改以及显示最终搜索到的结果信息。

Discusses how to solve the multi-objective network optimization by using the genetic algorithms （GA） to get the optimum solution or the almost optimum solution. In the design of the algorithms, the non-dominated Genetic Algorithm is used to solve the problem. The chromosome is coded according to the node-order on the path. Also, the length of the chromosome is the total number of the nodes. If a node is not on the path, the coding value in the chromosome code is zero. In the algorithms the tournament selection, partially matched crossover, exchanged mutation and elitist model to evolutes the individuals. The software is performed by the Visual C ＋ ＋. And the final results and some other information can also be showen in the windows, including the change of the relative parameters of the genetic algorithms.