论文部分内容阅读
提出一种基于人工免疫响应的线性系统逼近算法.给出了人工免疫响应的四元组模型,为免疫响应过程建立了一个可用于工程计算的数学模型;设计了克隆选择、免疫记忆和免疫调节等具体操作,模拟了抗体克隆选择、免疫记忆、基因免疫、免疫耐受等现象,实现了人工免疫响应的记忆学习.基于抗体群的随机状态转移过程,证明了新算法具有全局收敛性.基于两个典型的稳定或非稳定线性系统逼近问题的数值试验表明,无论在固定的区间内搜索还是在动态扩展的区间内搜索,人工免疫响应算法都能得到线性系统的最优逼近模型,算法是有效的.
A linear system approximation algorithm based on artificial immune response is proposed.The quaternion model of artificial immune response is given, a mathematical model which can be used for engineering calculation is built for immune response process, and the clonal selection, immune memory and immunomodulation And so on, the phenomena of clonal selection, immune memory, gene immunity and immune tolerance were simulated to realize the memory learning of artificial immune response.Based on the random state transition process of antibody population, the new algorithm is proved to be globally convergent. Numerical experiments on two typical linear approximations of stable or unstable linear systems show that the artificial immune response algorithm can get the optimal approximation model of linear system, both in a fixed interval search and in a dynamically expanding interval. The algorithm is Effective.