自适应迁移并行遗传算法在无线通信网优化中的应用

为利用最少的基站数实现规定的服务质量 ,提出了在一个区域的若干个备选位置中选择最优组合的粗粒度并行遗传算法 ,在不提高成本的同时 ,优化网络的覆盖率。不但提出了一种高效的局部搜索算子 ,而且为防止未成熟收敛 ,使用了只在子群体无法继续进化时才接收迁入个体的自适应迁移策略。仿真证明 ,对比固定周期迁移 ,采用的自适应迁移策略能够提高优化的速度和质量。

Radio network design using adaptive-migration parallel genetic algorithms

An adaptive -migration coarse -grained parallel genetic algorithm was used to choose transmitter locations from a set of possible sites in a known area which guarantees a stipulated quality of service (QOS). The algorthm maximizes the coverage while reducing the number of transmitters. An effective local search operator was developed, and a migration strategy, which a subpopulation accepted immigrants only when it could not evolve continuously by itself, was applied to avoid premature convergence. Simulations on a dedicated cluster demonstrate that relative to a fixed -epoch migrating algorithm, the adaptive -migration strategy improves the quality and the speed of optimization.