Modification of the Affine Scaling Interior-Point Algorithm for Multi-Objective Linear Programming

Abstract

This paper modifies Arbel’s affine scaling interior point multiple objective linear programming (ASIMOLP) algorithm to resolve zigzagging during the search process. This was achieve by developing a weight generating priority function rather that using the traditional AHP whose priority vector changes at each iteration; the newly developed weighting function generates equal weighting coefficients depending on the number of objectives in a problem which are then applied to the projected gradient that are produced by the algorithm. The modified ASIMOLP algorithm speed up convergence to the efficient frontier in an acceptable number of iterations thereby dramatically improving computational efficiency of the algorithm as seen in the numerical illustration.