一种新颖的函数优化进化算法

进化算法在求解全局优化问题时易陷入局部最优且收敛速度慢. 为了解决这一问题, 设计了一个基于下降尺度函数的杂交算子, 利用下降尺度函数与种群的关系来寻找实值函数的下降方向. 为了提高非均匀变异算子在进化后期的搜索能力, 通过均衡算子的局部搜索和全局搜索能力使其在算法后期仍能跳出局部最优. 在此基础上给出了一种新的进化算法. 最后将其与9个现有的算法进行了比较, 数值实验表明新算法快速有效.

A novel evolutionary algorithm for function optimization

When an evolutionary algorithm is applied to global optimization problems, it may be trapped around the local optima of the objective function and has a low convergence-rate. To solve these problems, a crossover operator is developed based on a descent-marking function. This operator finds descent directions based on the relation between the descent-marking function and the population. To improve the search ability of a non-uniform mutation operator in the late stage of evolution, an improved non-uniform mutation operator is designed for balancing the ability of global search and local exploration, which makes the algorithm able to avoid the premature convergence in the final stage of evolution. Combining all these techniques, we present a novel evolutionary algorithm. The presented algorithm is compared with 9 existing ones by simulations. Finally, experimental results indicate that the proposed algorithm is fast and efficient for all the test functions.