基于遗传算法的函数优化拟下降算法

Quasi Descent Algorithm for Function Optimization Based on Genetic Algorithm

在传统优化算法中嵌入一拟下降步，改善了算法的收敛性质；用遗传算法计算拟下降步，提高了算法的全局搜索能力，又不改变原算法的收敛性．数值计算结果表明，所提出的算法结合了遗传算法和传统算法的优点，是一种可行的解决函数优化问题的混合算法．

By adding a quasi descent step into the traditional optimization, the convergence of the algorithm is improved. The step is solved by the genetic algorithm, so that it not only enhances the global search ability of the traditional algorithm but also retains its convergence. The numerical computation shows that this method possesses the advantages of both the genetic algorithm and the traditional one. It is a feasible hybrid algorithm for function optimization.