Tactical fleet mix computation using multiobjective evolutionary optimization

A multi-objective model for determining optimal fleet mixes for a set of non-concurrent tactical missions is studied. An implementation of the multiobjective evolutionary algorithm (MEA) NSGA-II is used to solve the problem, and it is compared to an exhaustive search of the fleet mix space as well as a solution set obtained by a Mixed Integer Non-Linear Program. The exhaustive and MEA searches are compared with respect to both accuracy and computational complexity. The use of an alternative genotypic representation is investigated; its slow convergence is used to demonstrate the importance of chromosome choice, aligning with past research on locality and redundancy in genotypic representations.