A novel immune algorithm for complex optimization problems

A novel artificial immune algorithm using a self-pattern-recognition technique in vaccine extraction procedure is proposed for solving complex optimization problems. It takes objectives and constraints of an optimization problem as antigens, feasible solution of the problem as an antibody, and characteristic knowledge utilization as vaccine inoculation. By automatic recognition of the patterns in effective antibodies at each cultivation period, this algorithm boosts or restricts the generation of new antibodies in order to find the optimal solution more efficiently. Simulation study is conducted on the optimization of a series of benchmark problems in comparison with some other ecological algorithms- and colony algorithm and genetic algorithm. The statistical analysis results demonstrate that the algorithm proposed in this paper has better computation precision and searching efficiency in solving strong nonlinear optimization problems.