Hybrid Estimation of Distribution Algorithm for Multiobjective Knapsack Problem


We propose a hybrid estimation of distribution algorithm (MOHEDA) for solving the multiobjective 0/1 knapsack problem (MOKP). Local search based on weighted sum method is proposed, and random repair method (RRM) is used to handle the constraints. Moreover, for the purpose of diversity preservation, a new and fast clustering method, called stochastic clustering method (SCM), is also introduced for mixture-based modelling. The experimental results indicate that MO-HEDA outperforms several other state-of-the-art algorithms.