Coevolutionary optimization algorithm with dynamic sub-population size

This paper proposes a coeυolutionary optimization algorithm called DCOA. DCOA mainly focuses on how to adjust sub-population size self-adaptively so as to improve the optimizing performance. To achieve this, a strategy is introduced which consists of three rules: internal competition, external competition and spontaneous growth rules. These rules can control individual reproduction and elimination speed in each sub-population. Furthermore, the adjustment can be proven globally asymptotically stable. In the experiments, we compare the performances of DCOA, macroevolutionary algorithm (MA) [13] and simple genetic algorithm (SGA) with typical test functions. The results show that DCOA is able to find the global optimum on most difficult functions, nothing less than MA which uses simulated annealing technique. At the same time, DCOA converges quickly, similar to SGA and faster than MA.