论文部分内容阅读
提出一种定向多尺度变异克隆选择优化算法.为了实现抗体间信息共享,算法利用定向进化机制引导抗体向着抗体群最优解区域逼近.采用多尺度高斯变异机制,在算法初期利用大尺度振荡变异实现了全局最优解空间的快速定位.随着适应值的提升,小尺度变异会随之减低,使得算法在进化后期通过小尺度变异完成局部精确解的搜索.将算法应用到5个经典函数优化问题,结果表明,该算法不仅具有更快的收敛速度,而且全局解搜索能力和稳定性均有显著提高.
In order to realize information sharing between antibodies, the algorithm uses directed evolution mechanism to guide the antibody towards the optimal solution region of the antibody population.Using multi-scale Gaussian mutation mechanism, the large-scale oscillatory variation The global optimal solution space is rapidly located, and as the fitness value increases, the small-scale variation will decrease, which makes the algorithm complete the local exact solution search through small-scale variation in the late evolutionary stage.The algorithm is applied to five classical functions The results show that the proposed algorithm not only has faster convergence speed but also significantly improves the global search capability and stability.