一种改进的混合遗传算法研究

标准遗传算法的求泛能力优于它的求精能力，在求解GA-困难问题时求解精度难以控制，本文由此提出了一种改进的ε-混合遗传算法。本算法在每代找出最优个体之后，以该最优个体为初始出发点在一个固定半径的区域内进行局部搜索，以搜索结果代替最差个体或其它个体，然后再进入下一代操作。算法大大提高了求解精度，同时也提高了稳定定性。

Research of an Improved Hybrid Genetic Algorithm

The extensive ability of SGA surpasses its refinement ability. It is difficult to control the precision when we solve the GA-hard problems. This paper proposes a new hybrid genetic algorithm,ε-HGAA. After finding the most optimal individual in each generation, it takes this optimal individual as th initial point to carry out local search in a fixed-radius region. The search results replace the worst individual or other individuals; and the algorithm enters the next generation operation again. This algorithm not only increases the solution precision, but also enhances the stability.