This paper describes a scalable algorithm for solving multiobjectivedecomposable problems by combining the hierarchical Bayesian optimizationalgorithm (hBOA) with the nondominated sorting genetic algorithm (NSGA-II) andclustering in the objective space. It is first argued that for goodscalability, clustering or some other form of niching in the objective space isnecessary and the size of each niche should be approximately equal.Multiobjective hBOA (mohBOA) is then described that combines hBOA, NSGA-II andclustering in the objective space. The algorithm mohBOA differs from themultiobjective variants of BOA and hBOA proposed in the past by includingclustering in the objective space and allocating an approximately equally sizedportion of the population to each cluster. The algorithm mohBOA is shown toscale up well on a number of problems on which standard multiobjectiveevolutionary algorithms perform poorly.
展开▼