具有捕食策略的混合随机优化算法及其多极值函数优化

Integrated algorithm of simulated annealing, genetic algorithm and chemotaxis algorithm with predatory search strategy and its application in multi-valued function optimization

SAGACIA是一种混合随机优化算法,该算法虽已吸收了模拟退火算法、遗传算法和趋化性算法的优点,但搜索过程中仍存在收敛速度慢以及采用固定步长影响搜索精度的缺点,而捕食搜索策略通过限制的调节能较快锁定最优区域,从而提高收敛速度。结合两者的优缺点,提出一种具有捕食搜索策略的自适应调整步长SAGACIA算法,改进后的算法通过捕食搜索策略平衡了算法的局域搜索和全局搜索,提高了收敛速度;邻域搜索采用自适应步长,避免了最优解附近的震荡,提高了搜索精度。实验仿真结果表明,改进后的SAGACIA算法具有较快的收敛速度和较高的寻优精度,证明了算法改进的有效性和可行性。

The Integrated Algorithm of Simulated Annealing, Genetic Algorithm and Chemotaxis Algorithm（ SAGACIA）is a hybrid stochastic algorithm, which absorbs advantages of simulated annealing algorithm, genetic algorithm and chemotaxis algorithm. But slow convergence and bad search accuracy with fixed step in searching still exist. Predator Search strategy（ PS）can look for better region fast through adjusting restriction. Combining their advantages and disadvantages, a self-adaptive step SAGACIA optimization algorithm was proposed based on predatory search strategy. Local search and global search were balanced better through predatory search strategy, so convergence speed was improved. Adaptive step was used in neighborhood search, which avoided shock near optimal solution and improved search accuracy. Simulation results show that the convergence speed and search accuracy of the improved SAGACIA are improved. The improved algorithm is a feasible and effective method.