An improved algorithm exploiting the characteristics of a distance-based preference function to converge to preferred solutions

We address the problem of choosing the most preferred of a set of alternatives that are defined by multiple criteria. We assume that the decision maker’s preferences can be represented by a general class of weighted distance functions that can take a wide variety of forms. We exploit the characteristics of these functions and develop an interactive algorithm that guarantees to find the most preferred alternative of a decision maker whose preferences are consistent with a distance-based function. In contrast with a benchmark algorithm that uses similar preference functions, our algorithm moves through different distance functions effectively to converge to the best alternative quickly. Our experiments on a variety of three- and four-objective problems demonstrate that our algorithm performs well, far outperforming the benchmark algorithm in terms of the required preference information from the decision maker.