Multiobjective evolutionary algorithm with risk minimization applied to a fleet mix problem

Abstract

We apply the non-dominated sorting genetic algorithm-II (NSGA-II) to a multi-objective fleet-mix problem for risk mitigation. The Stochastic Fleet Estimation (SaFE) model, a Monte Carlo-based model, is used to determine average annual requirements which a fleet must meet. We search for Pareto-optimal combinations of platform-to-task assignments that can be used to complete SaFE generated scenarios. Solutions are evaluated using three objectives, with a goal of minimizing fleet cost, total task duration, and the risk that a solution will not be able to accomplish future scenarios. Optimization over all three objectives allowed for exploration of configurations which were low cost and low risk, a region not explored by prior experiments without the risk objective.