A Modular Decomposition for Approximation of the Optimum DSS in Multidimensional Cases

Abstract

Decision making tasks are often complex and ill-defined, and they are also multi-aspect because many factors influence the decision process. Decision support systems (DSS) which are frequently used in business and manufacturing have a potential to cope with vagueness and ill-structure of complex decision making situations. Also an optimum DSS can be developed for multi-dimensional decision problems. Complexity of such a system however rapidly grows with an increasing number of influencing factors, their dimensions and their relationships. As a consequence development, operation and maintenance of the optimum DSS becomes difficult and sometimes even unfeasible for real situations. This paper describes a feasible approximation of the optimum DSS through decreasing its complexity by a modular decomposition. The resulting DSS consists of multiple DSSs, each of them is focused at the individual decision making dimension. Paper also presents a practical simple integration mechanism and heuristic for integrating partial DSSs.