Adjusting Trade-Offs in Multi-Criteria Decision-Making Problems

Abstract

In multi-criteria decision-making (MCDM) methods, allowing trade-offs between the decision criteria may result in inappropriate conclusions. Allowing trade-offs implies that the weaknesses of an alternative in some criteria may be compensated by its strengths in other criteria. While, with some problems, there are no concerns with regard to allowing a full trade-off, in other cases, such a trade-off may not follow the actual decision-making problem. This paper proposes a new approach based on defining an upper and lower expectation level for each decision-making criterion. Indeed, the proposed approach is a framework by which the requirements of a decision-maker are considered in the criteria involved in MCDM problems. That distinction generates two primary and secondary performance matrices. The primary matrix includes the values of the alternatives with respect to the individual criterion up to the upper expectation levels of the decision-maker, while the secondary matrix, which is defined by the amounts above the upper levels and below the lower levels of the decision-maker’s expectations, contains each alternative in all the criteria and, to some extent, can exceed the upper levels of the decision-maker’s expectations. The final result of each alternative is calculated by adding its outcome in the two matrices. The results of a numerical example involving cellphone selection show how controlling the trade-offs could affect the results.