An improved particle swarm optimization algorithm based on comparative judgment
This study proposes a modified artificial immune network algorithm for function optimization problems based on idiotypic immune network theory. A hyper-cubic mutation operator was introduced to reduce the heavy computational cost of the traditional opt-AINet algorithm. Moreover, the new symmetrical mutation can effectively improve local search. To maintain population diversity, we also devised an immune selection mechanism based on density and fitness. The global convergence of the algorithm was deduced through the method of pure probability and iterative formula. Simulation results of benchmark function optimization show that the modified algorithm converges more effectively than other immune network algorithms.