Closest reasonable centroids: A new approach for MADA problems

Abstract Weight has been widely used in multi-attribute decision analysis (MADA) problems. However there exist several challenges of using weight mechanisms. This paper devises a weightless mechanism, called closest reasonable centroids (CRC), for MADA problems, in particular suitable for the cases that weight info rmation is either unavailable or unattainable. The proposed mechanism aims to make a feasible guess about the solutions for the MADA problems through two stages. Firstly, the attribute scores given by evaluators are aggregated into centroids. Secondly, genetic algorithm (GA) is then employed to provide a feasible search of the search space of all the possible combinations of centroids. Experimental results show that CRC shows a satisfactory result in many decision contexts, when compared to a number of other traditional weighting approaches.