An essential feature of active Decision Support Systems (DSS) is the ability to take the initiative in performing decision-related tasks. One possibility for providing active high cognitive level decision support is through facilitating alternative generation in DSS. The method proposed in this work enables the generation of several diverse alternatives in a single run. The method relies on the principles of effective problem-solving/decision-making and facilitates divergent processes, the separation of alternative generation from evaluation, as well as the diminishing of human cognitive biases. A hybrid DSS based on genetic algorithms (GA) and fuzzy sets is used to operationalize the approach. The paper outlines the design requirements for alternative generation in DSS and discusses the inadequacies of the "what-if" simulation and traditional optimization methods in light of these requirements. The paper further elaborates on the appropriateness of GA as a tool for alternative generation in DSS for solving complex ill-structured problems. The method is illustrated using marketing mix problem in a simulated business environment. The results suggest that the GA-based alternative generation leads to promising diverse alternatives. An active DSS incorporating the proposed method reduces the time-consuming manual search for promising alternatives and provides a higher degree of man-machine collaboration.
Determining user requirements and generating alternative system solutions to meet these requirements are two critical steps in the requirement analysis phase of the system development life cycle. Much of the MIS research in the requirement analysis phase has been devoted to the topic of requirement determination and its verification. Alternative generation and evaluation is left, to a significant degree, to the judgment and expertise of an analyst. This paper proposes a multiple criteria decision making (MCDM) approach for generating and evaluating alternatives when the user requirements are expressed in terms of certain operational criteria such as time, cost, risk, etc. These alternatives form the basis for the user to make the necessary trade-offs.