Preference-based Monte Carlo weight assignment for multiple-criteria decision making in defense planning

In this paper we develop a multiple process approach for finding the most preferred plan instance within military defense planning. In this problem tens of thousands of alternative plans with different simulation settings are evaluated by multiple measures of effectiveness through simulation. In a sequence of processes, the Pareto optimal frontier of the plans is first found. Using decision makers' preferences where the decision makers may provide incomplete preferences regarding subsets of disjoint measures, these measures are ordered by their importance. Alternative sets of random weights are assigned to the set of measures by a Monte Carlo approach where weights uphold the preferred order. Finally, all plans are evaluated using the weighted measures and the most preferred plans according to a ranking index are further analyzed to provide decision support.