Aggregation of dependent criteria in multicriteria decision making problems by means of capacities

We address the multiple criteria decision problem in which the criteria are mutually dependent, in the context of land combat vehicle selection for Australian Defense. We employ the theory of capacities to model criteria dependencies. Its challenge is exponentially growing (with the number of criteria) set of parameters that represent all criteria interactions. We outline a number of strategies to (a) simplify the construction process and reduce the number of parameters, and (b) elicit preferences from the decision makers and translate them into capacity learning problem. We present various challenges and ways to address them from the mathematical programming perspective.