求解高维复杂函数的遗传-灰狼混合算法

Hybrid genetic grey wolf algorithm for high dimensional complex function optimization

高维函数优化一般是指维数超过100维的函数优化问题,由于"维数灾难"的存在,求解起来十分困难.针对灰狼算法迭代后期收敛速度慢,求解高维函数易陷入局部最优的缺点,在基本灰狼算法中引入3种遗传算子,提出一种遗传-灰狼混合算法(hybrid genetic grey wolf algorithm,HGGWA).混合算法能够充分发挥两种算法各自的优势,提高算法的全局收敛性,针对精英个体的变异操作有效防止算法陷入局部最优值.通过13个标准测试函数和10个高维测试函数验证算法的性能,并将优化结果与PSO、GSA、GWO三种基本算法以及9种改进算法进行比较.仿真结果表明,所提算法在收敛精度方面得到了极大改进,验证了HGGWA算法求解高维函数的有效性.

High-dimensional function optimization usually refers to the function optimization problem with dimension over 100,which is difficult to be solved for the existence of"dimension disaster".In this paper,three genetic operators are embedded into the basic grey wolf algorithm,and a hybrid genetic-grey wolf algorithm(HGGWA)is proposed.The global convergence of the HGGWA is greatly improved by combining the advantages of the GWO and GA.The current three optimal individuals are disturbed by the diversity mutation operator in the process of the search so as to avoid the possibility of falling into local optimum.The performance of the algorithm is verified using 13 standard benchmark functions and 10 high dimensional functions,and the optimization results are compared with the PSO、GSA、GWO and 9 improved algorithms.Simulation results show that the HGGWA is greatly improved in convergence accuracy,which verify the effectiveness of the HGGWA in solving high-dimensional functions.