A New Multi-modal Immune Algorithm and its Complete Convergence Analysis

Many available multi-modal optimization algorithms have good search effects on functions with equivalence peaks, but they cannot find all extremum points in the definition fields, i.e. they are not complete convergence. So a strategy of peaks poised is proposed and a new Multi-modal Immune Algorithm (MIA) based on the antibody evolution in immune system with this strategy is designed. Then the algorithm´s complete convergence and its computation complicacy are analyzed. Simulation experimental results testify that MIA has excellent availability on finding all extremum points of a multi-modal functions, especially those functions with non-equivalence peaks.