改进的混合级联遗传算法

Improved Hybrid Cascaded Genetic Algorithm

将模拟退火算法和级联遗传算法相结合,提出了一种改进的混合级联遗传算法。模拟退火算法承认物种进化过程中的局部失败和倒退,它允许进化中的波折而不是非要物种进化一直是上升的、成功的,模拟退火算法能使搜索过程避免陷入局部最优解。级联遗传算法假设问题的最优解总是靠近于问题的局部最优解的,据此,级联遗传算法通过不断缩小解空间达到快速收敛的目的。综合运用这两种算法,可克服模拟退火算法收敛速度慢、级联遗传算法局部搜索能力差的缺点。利用本算法构造CL多小波前置滤波器的实验结果表明,本算法不仅计算速度快,而且稳定性也得到了显著提高。

Combined the SAGA （Simulated Annealing Genetic Algorithm） with CGA （Cascaded Genetic Arithmetic）, an improved hybrid cascaded genetic arithmetic is proposed. The SAGA accepts the partial loss and devolution in the course of species＇ evolution, and allows the fluctuate rather than the constant rise or success in evolution, it can make the course of search avoid being plunged into the optimized solution. The CGA supposes that the optimized solution is always close to the partial optimized solution, thus, the CGA achieves the rapid convergence by shrinking the space of solution. Using the two arithmetic synthetically, the shortcoming of slow convergence for SAGA and less capability of partial search of CGA is overcome. The results of the experiment using this arithmetic to construct CL multi-wavelet pre-fiher show that not only the compute speed is fast, but also the stability is remarkable improved.