论文部分内容阅读
针对光网络时延受限光组播路由计算复杂度高的问题,提出一种基于改进的和声搜索求解时延受限光组播路由的算法。论文通过引入自适应的和声记忆保留率及微调概率提高和声算法搜寻到全局最优光组播树的能力;算法设计了一个能够保持备选光组播树多样性的动态适应度函数以扩大光组播路由的搜索范围;同时,算法在初始化和声库时采用精英保留策略简化和声迭代过程。仿真结果表明,提出的改进算法能够求得代价更低的光组播路由,且算法在较大光网络规模下具有较好的收敛稳定性。
Aiming at the problem of high computational complexity of optical multicast routing with limited multicast delay, this paper proposes an algorithm based on improved harmonic search to solve constrained optical multicast routing with delay. In this dissertation, adaptive harmony memory retention rate and fine tuning probability are used to search the global optimal optical multicast tree. The algorithm designs a dynamic fitness function that can maintain the diversity of alternative optical multicast trees Expand the search scope of optical multicast routing; meanwhile, the algorithm adopts the elite retention strategy to simplify the harmony iteration process when initializing the sound library. The simulation results show that the proposed improved algorithm can find the optical multicast routing with lower cost, and the algorithm has better convergence stability under larger optical network scale.