Comparison study of arbitrarily-chosen weights vs optimized weights in multi-objective geometric programming

摘要

This paper provides a study into the impact of weight factors selection in a multi-objective geometric programming (MOGP) optimization. Typically, when solving a multi-objective design problem, a designer would arbitrarily assign weight factors based on his/her perception of each objective's significance, e.g. assigning higher weights to the more important objectives. Here, an alternative method is proposed to take into account of weight factors as additional optimizing variables and then formulate the new MOGP problem. Comparing the conventional and the proposed method, the latter can achieve a closer solution to the ideal multi-objective optimum by considering the weighted deviation sum between the two schemes.