Evaluation and refinement of aspiration-based methods in MCDM

The performance of aspiration-based interactive methods of MCDM is evaluated by means of a simulation model, based on idealized assumptions regarding preferences of the decision maker (DM). It is demonstrated that naive application of these procedures can lead to premature termination of the procedure at relatively poor solutions (i.e. relative to the idealized preferences), leaving much of the decision space unexplored by the DM. This suggests practical steps to build a form of ‘intelligence’ into algorithms for applying aspiration-based methods, to ensure that an improving sequence of solutions is examined, and that early termination is avoided. The simulation studies show that with the implementation of these steps, aspiration-based and value function-based methodologies produce comparable results, with the aspiration-based methods showing greater robustness to violations of preferential independence of the criteria.