Introducing niching into algorithm could maintain the diversity of results, also have better ability of global optimization and convergence rate, and is more adopted louse in optimization problem of multimodal function. In older to improve the efficiency of generating detectors and reduce the cost of algorilhm.it uses niching evolutionary theory in the immature detectors generation. According to the hamming similarity between the strings of initial population and the affinity and fitness of strings, niches are built and evolved respectively, an detector generation algorithm based on niching strategy is proposed. The parameters which involved in simulation experiment are discussed; The algorithm compares with exhaustive detector generation algorithm. Through analysis of experimental results, the algorithm basically reached the expected goal under certain conditions.%引入小生境思想能够较好地维持解的多样性,同时具有较高的全局寻优能力和收敛速度,特别适合于复杂多峰函数的优化问题.为了提高检测器的生成效率,降低算法开销,引入小生境进化原理用于未成熟检测器的生成;通过计算初始种群个体之间的海明相似度,依据个体之间的亲和力和适应度,构建多个小生境并分别进行进化,提出了基于小生境策略的检测器生成算法;对算法进行了验证,讨论了实验中涉及的参数,与穷举检测器生成算法进行了比较;通过对实验结果的分析,在一定条件下,该算法基本达到了预期的目标.
展开▼