论文部分内容阅读
针对进化算法随机盲目搜索的缺点,提出一种新的自适应梯度信息指导交叉的进化算法.该算法首先利用混沌序列初始化种群,在迭代过程中,根据当前最优个体的梯度信息和种群与个体的聚集程度,自适应地确定最优个体的负梯度方向范围,在该范围内随机选择个体与当前最优个体进行算术交叉操作,使交叉后的个体以较大概率向较好解的方向进化.另外,引入自适应变异算子用于平衡算法的开发和探测能力.几个典型测试函数的实验结果表明,新算法具有较高的收敛精度.
In order to overcome the shortcomings of random blind search of evolutionary algorithms, a new evolutionary adaptive gradient information guidance algorithm is proposed, which firstly initializes the population by using chaotic sequences. In the iterative process, according to the current optimal individual’s gradient information and population and individual , The negative gradient direction range of the optimal individual is adaptively determined. Within this range, the randomly selected individuals and the current best individual are subjected to an arithmetic crossover operation, so that the crossed individuals will evolve toward a better solution with greater probability In addition, the adaptive mutation operator is introduced to balance the development and detection of the algorithm.The experimental results of several typical test functions show that the new algorithm has higher convergence accuracy.