Currently, Genetic Algorithm (GA) has been applied in many practices. Some major drawbacks of GA such as premature, sticks to local optimum easily, low local search ability and slow speed of convergence, have been found. The two drawbacks, premature and sticks to local optimum easily, are mainly related to the selection pressure of GA and can be overcome by using the technique of fitness scaling, which is usually problem-dependent.The essential difference between GA and Artificial Immune Algorithm (AIA) is that AIA mimics the antibody reproduction strategy in natural immune system and employs concentration regulation mechanism, which makes the antibody with high fitness and low concentration proliferate and in the mean time surpasses the antibody with high concentration. Thus, AIA can regulate the selection pressures efficiently and keep the diversities of solution set and overcome the drawbacks existent in GA such as premature and local optimum. On the other hand, there exist drawbacks in AIA such as slow running and convergence speed.In this paper, the author tries to develop new methods to solve the aforementioned drawbacks in GA and AIA. The followings are the creative achievements the author achieved.1. In order to improve the local search ability of GA, a king crossover strategy was proposed. Combining elitist GA (EGA) with the king crossover strategy, a new elitist GA based on the king crossover strategy (KEGA) was developed. The global convergence of KEGA was proved. The result from optimizing F6 function (Schafferl function) using KEGA demonstrates that the on-line performance, off-line performance, optimum solution search ability of KEGA are better than that of EGA, the number of average generation which KEGA needed to converge to a solution is only one ninth of that of the EGA. The same result can also be achieved when optimizing F8 function (Rastrigin's function). Comparing to EGA, KEGA is easy to realize and the increase of computation load can be neglected, but the performance of KEGA can be improved greatly.2. To enhance the speed of AIA, two new methods of computing the concentration of antibody were developed. Simulation results for F8 and F15 (needle in haystack: type I) functions show that the two methods can improve the running... |