针对标准遗传算法在优化应用中遇到的诸如局部搜索能力差、计算量大、对较大搜索空间适应能力差和早熟收敛等问题,该文通过将免疫算法引入到遗传算法中,利用免疫算法的免疫记忆、自我调节和多样性保持功能弥补其不足,提出了一种基于信息熵的DNA免疫遗传算法.该算法采用DNA链对抗体进行编码,利用信息熵来表示抗体间的亲和度及浓度,并提出了一种新的评估指标--聚合亲和度,有效地实现了抗体群的自我调节和多样性保持策略.最后,利用典型测试函数验证了本文方法的有效性.
In this paper, by incorporating immune algorithm's functions such as immune memory, self - adjustability and the keeping of the diversity into genetic algorithm, a novel DNA - immune - genetic algorithm based on information entropy is presented to overcome the shortcomings of standard genetic algorithm, e. g. , poor local search ability, premature convergence, excessive computational cost and bad adaptability to large search space, etc. The proposed algorithm codes the antibody by utilizing DNA chain, in which information entropy is introduced to denote the consistence and the affinity of the antibodies. As a novel evaluative index, the converged affinity is presented to achieve the self - adjustability of colony and keep the diversity of colony. Finally, the performance of the present algorithm is validated by means of typical test functions.