An Improved Estimation of Distribution Algorithm in Dynamic Environments

In dynamic environments, the optimal solution changes over time. To track the solution, an improved univariate marginal distribution algorithm (UMDA) is proposed. A transfer model is introduced to increase the diversity of population. The current information is used to avoid being trapped into the local optimization for dynamic optimization problems. The scheme is illustrated through simulations applying dynamic moving peaks benchmark. The results show that the proposed algorithm is effective and can accommodate the dynamic environments rapidly.