Decision making in complex environments is influenced by a large number of conflicting and incommensurable factors. It is multidimensional and the influencing factors affect the decision process to varying degrees. The performance measurement of decision alternatives in multi-criteria decision making (MCDM) may be represented as a multidimensional vector in real space. The technique for order preference by similarity to ideal solution (TOPSIS) is a popular MCDM tool. However, sometimes this technique shows its inability to differentiate between decision alternatives. In this paper the limitations of TOPSIS are examined, and an improvement is suggested. The proposed new method uses the vector dissimilarity approach to remove the boundary restrictions and achieve greater sovereignty. The improved method has been demonstrated through a case study of maintenance design for dump trucks deployed in a large surface mine.