一种基于小生境的混合遗传退火算法

分析遗传算法和模拟退火算法混合优化策略的构造出发点 ,融合小生境技术的思想 ,提出一种以遗传算法和模拟退火算法为子算法的基于小生境技术的混合遗传退火算法———NGSA算法 ,并对该算法的特点和优化性能作了定性分析。结合典型多峰值测试函数———Shubert函数的求解实验 ,说明NGSA算法具有较强的全局和局部搜索能力 ,能够高效地寻找到多个全局极值 ,且参数选择不必过分严格 ,是一种优化能力、效率和可靠性较高的多峰值优化方法。最后 ,讨论了该算法在机械学科的广泛应用背景。

A Niching Hybrid Genetic Simulated Annealing Algorithm

After analyzing the construction foundation for hybrid optimization strategy of genetic algorithm and simulated annealing algorithm, a niching genetic simulated annealing (NGSA) algorithm is presented, which is based on niche technology. The features and optimizing performances of NGSA algorithm are discussed. Shubert function, a representative multi-modal optimization problem, is used to verify the algorithm. The result shows that NGSA algorithm has a strong capability in global and local search, it can find all extrema in a short time without strict requests for parameters, so it is a good optimizing algorithm for multi-modal problems with higher capability, efficiency and reliability.