基于素描的遗传算法

Genetic Algorithm Based on Sketching

传统遗传算法容易陷入局部最优解,本文借鉴美术中“素描”的思想,对传统的遗传算法进行了改进,提出了基于素描的新型遗传算法.该算法模拟人的素描行为,构造参数控制下的选择算子,再通过参数的调节来选择个体,并依据最优个体对选择算子进行修正,以达到动态调整群体进化过程中的种群多样性和收敛速度之间的矛盾,从而有效地避免了传统遗传算法中早熟现象,显著地提高了GA对全局最优解的搜索能力和收敛速度.这将使GA在众多实际的优化问题上将具有更广泛的应用前景.仿真结果表明,该算法正确有效,且性能优于现有的其它方法.

An improved genetic algorithm based on the sketching in art is proposed to avoid the problem of local optimum. The key to this algorithm lies in the construction of a new selection operator controlled by a factor based on a simulation of human sketching behaviour, which selects the individuals by setting a varying factor and is changed according to the best individual of the new population. Therefore, the operator can reach dynamic adjusting the contradiction between the diversity of population and the convergent speed in the process of population evolution. Hence this new algorithm can avoid the premature convergence. Moreover, this algorithm improves the ability of searching an optimum solution and increases the convergence speed.This algorithm has extensive applications for many practical optimization problems. The simulation proved its effectiveness and better performance.