A Nonlinear Dynamic Method for Supporting Large-Scale Decision-Making in Uncertain Environments

This research developed a methodology for supporting decision making by reducing uncertainty in decision environments which are too large, dynamic and complex to be treated by traditional quantitative and simulation techniques. These environments are complex because of the free choice associated with human involvement, and the existence of a large number of interrelated factors which influence the outcomes of the decision process. They are dynamic because the ground rules affecting those interrelationships are constantly changing. Uncertainty cannot be treated probabilistically, since identification of a full set of outcomes and factors of influence is not possible. The venue for the investigation was the infrastructure which supports commercial space launch activities in the United States. The issue treated was whether it would be advisable to make large capital investment in that infrastructure. The problem was approached using the principles of Chaos Theory and Nonlinear Dynamics, in a manner similar to that used by Priesmeyer (1992). The intent was to engender a more systemic view of the environment and approach analysis by examining marginal changes, over a period of ten years, in factors which tend to influence the outcome. The objective was to develop hypotheses which, when validated, will provide a new perspective for decision makers from which to enhance the robustness of these kinds of decisions. The methodology, which evolved over several years of preliminary research, involved identification of sectors of the commercial space infrastructure, isolation of the more important decision factors, identification and solicitation of knowledgeable respondents from the various infrastructure sectors, development of a computerized qualitative data gathering instrument, and graphical analysis of data represented by phase plane diagrams. Although there was little evidence of classical chaotic behavior in the data, the analysis was able to isolate those nonlinear dynamic relationships between decision factors which appeared most likely to provide information regarding system behavior. One hypothesis was developed directly from that observation. A second resulted from the development of an aggregate measure of the level of uncertainty (and, consequently, investment risk) inherent in the decision environment