Immune Algorithm And Its Applications In Digital Filter Design

In this thesis we present two improvements in immune algorithm according to immune mechanism in the natural immune system after we analyzed the present realizations of immune algorithm. (l)We adopt a new process method with memory cell. It is more similar with the natural immune system by changing from storing one memory cell to storing a group of memory cells in order to improve the diversity in the colony. (2)We introduce the conception of vaccine by distilling the information of parents. We can gain an efficient crossover result by using the vaccine-based policy. To verify the above novel ideas, we analyze and implement the improved immune algorithm in the resolve of TSP. At last, we apply the improved immune algorithm to the design of LMS adaptive filter and FIR filter. In the first case we gained optimal value of the key parameter μ in LMS. In the second case, based on the frequency sampling technology we gained an ideal design result of FIR. To prove our immune algorithm' s excellent performance we compared the above results with other methods independently.