摘要
针对紧致遗传算法求解复杂优化问题的局限性,提出一种改进的高选择压力紧致遗传算法。该算法利用概率向量随机产生S(S>2)个个体,并按照适应度值进行排序,然后由最优解与其他解线性组合构成的虚拟解进行相互竞争,从而实现概率向量的更新。对3种不同类型测试函数的仿真结果表明,改进算法比标准紧致遗传算法和高选择压力紧致遗传算法具有更高的优化精度。
In order to improve the performance of the compact Genetic Algorithm（cGA） to solve more complicated optimal problems, an improved cGA with higher selection pressures is proposed. In the proposed algorithm, S（S〉2） individuals are generated from the probability vector, and then the best individual selected as the winner is obtained by their ranking order of the fitness value. In the competition process, the winner solution competes with the virtual solution composed of the linear combination of the other S-1 solutions. The probability vector is then updated towards the winner until the probability vector is converged. Experimental results on three different kinds of benchmark functions show that the proposed algorithm has higher precision of optimization than that of the standard cGA and the cGA with higher selection pressures.
出处
《计算机工程》
CAS
CSCD
北大核心
2008年第24期183-185,共3页
Computer Engineering
关键词
分布估计算法
紧致遗传算法
选择压力
estimation of distribution algorithms
compact Genetic Algorithm（cGA）
selection pressures
