Evolutionary algorithm using converted problems

Any evolutionary algorithms should conduct biased search in the search space. Most popular strategies for doing so focus on how to select and generate solutions. In this paper, a new strategy is proposed. It completes the task through the transformation of the given problem. A population of converted problem responding various search weights is used, which may be more suitable for evolutionary algorithm to solve. To show the performance of the new strategy, instantiated algorithm is designed. On some trap problems and benchmark problems, the proposed algorithm using converted problems has competitive performance.