MASP: A Multi-attribute Secretary Problem Approach to Multi-objective Optimization with Fair Decision Maker
Mario Köppen, Rodrigo Verschae, Masato Tsuru · 2012
In the present paper we propose a multi-objective optimization procedure inspired by the famous secretary problem from optimal stopping theory. the proposed algorithm is designed to be applied to multi-attribute decision making problems that require fair solutions to be obtained. We consider two fairness relations, namely maxmin fairness and proportional fairness, and the evaluation is performed on the problem of wireless channel allocation. the performance is compared to random search, and the quality of the solution is measured with respect to the distance to the results obtained with brute-force in the whole search space. Results show that the proposed algorithm, compared to random search, can cover a much larger portion of the search space for the same number of pair wise comparisons.