【摘要】 针对克隆选择算法自适应能力和多值搜索能力较弱的不足,提出了一种基于危险理论的自适应免疫算法。算法中引入种群环境和抗体危险信号引导自适应免疫应答过程,增强了种群多样性,避免了算法过早收敛。利用Markov链证明了算法的收敛性,分析了算法的复杂度。针对经典benchmark函数的仿真实验结果表明,相比克隆选择算法,本算法具有良好的全局收敛能力和多值搜索能力,且具备较快的收敛速度和求解精度。更多还原

【Abstract】 Since clonal selection algorithm lacks of adaptive capacity when solving multimodal problems,a novel adaptive immune algorithm,which was based on immune danger theory,immune network and clonal selection theory,was proposed to emulate the entire immune mechanisms and to enhance the performance for complex multimodal problems.The environment of antibody population and the corresponding danger signal of each antibody were incorporated into the process of immune response,which preserved the diversity of antibody population and then alleviated the premature of the algorithm to some extent.The algorithm was proved theoretically to be convergent with Markov chain model.Simulation results on the classical benchmark functions showed that,compared to clonal selection algorithm,this algorithm has good performance of global convergence and multimodal searching ability,and has a fast convergence speed with good quality of solution.更多还原