Towards a small diverse pareto-optimal solutions set generator for multiobjective optimization problems.

Multiobjective evolutionary algorithms (MOEAs) try to produce enough and sufficiently diverse Pareto-optimal tradeoff solutions to cover the entire Pareto surface. However, in practical scenarios, presenting numerous solutions to stakeholders may result in confusion and indecision. This paper proposes a method for generating a small (user-specified) number of well-distributed Pareto-optimal feasible solutions for multiobjective problems. The proposed method can be applied to a set of aggregate solutions produced by (1) one MOEA over multiple runs, (2) several different MOEAs, or (3) a universal set of feasible solutions produced by one or more constraint solvers.