High-Dimensional Real Parameter Clonal Selection Memory Algorithm

For numerical high-dimensional real parameter optimization problem, a new clonal selection memory algorithm is proposed in the paper by introducing a short-term memory mechanism into the optimization algorithm. Through non-genetic information storage and guiding the subsequent evolution the algorithm can effectively increase the local convergence rate. On the other hand, the search depth threshold setting, supplement operator and crossover operator makes the algorithm can effectively escape local optimum traps. Combined with other intelligent algorithms, the simulation experiment of the 20 and 30 dimensional standard test functions show that the new algorithm has obvious advantages in convergence speed, convergence precision and global convergence. Furthermore, the simulation experiment of the 100 and 200 dimensional standard test functions shows that, the new algorithm shows better global convergence speed, convergence precision and stability.