利用遗传算法实现基于多目标约束的网络规划

摘要

Computer network topology planning,with 2 -terminal network reliability and all -terminal network connectivity,which led by communication backbone network with network topology cost,node load and reliability restriction,is a multi -objective optimization problem.By defining and studying a hybrid network reliability planning model,the global preferential random selection strategy, random multi -point crossover and mutation strategy are used to improve the genetic algorithm for network topology planning under multi objects restriction.The test result shows that the feasibility of study is provided and the efficiency of the network topology optimization is improved.%计算机网络拓扑规划通常是在通信主干网主导下受网络拓扑成本、节点负载、可靠性约束等制约的2－终端网络可靠性与全终端网络连通性混合的多目标优化问题。通过定义并研究这一混合网络的可靠性规划模型，采用全局择优随机选择策略以及随机多点交叉和变异策略改进遗传算法进行求解，实现了多目标约束下的网络拓扑规划。测试软件证明了研究的可行性并有效提高了网络拓扑布局的优化效率。