An effective and efficient evolutionary algorithm for many-objective optimization

In evolutionary multiobjective optimization, effectiveness refers to how an algorithm performs in terms of converging its solutions into the Pareto front and also diversifying them over front. This is not easy job, particularly for optimization problems with more than three objectives, dubbed many-objective problems. such problems, classic Pareto-based algorithms fail provide sufficient selection pressure towards front, whilst recently developed algorithms, as decomposition-based ones, may struggle maintain a set well-distributed on certain (e.g., those irregular fronts). Another issue some optimizers rapidly increasing computational requirement number hypervolume-based shift-based density estimation (SDE) methods. this paper, we aim address problem develop effective efficient (E3A) that can handle various E3A, inspired by SDE, novel population maintenance method proposed select high-quality environmental procedure. We conduct extensive experiments show E3A better 11 state-of-the-art quickly finding well-converged well-diversified solutions.