Abstract :
Forecasting, in either implicit or explicit form, is a crucial capability for information-decision-action systems operating in uncertain environments. Strategic and operational decision making may depend heavily on predictions of what the physical environment, markets, competitors or enemies will do. However, the forecasting methods which are available at present divide between quantitative techniques and qualitative methods. Since neither technique can do the job of the other, and most problem domains present a mixture of quantitative and qualitative problems, this jeopardises the creation of a coherent management and control system and produces a ragged interface between technological decision support and human decision makers. The techniques given for structuring, performing, improving, evaluating and (eventually) automating qualitative forecasting tasks, can help to address these issues in several ways: by extracting and explicating the predictive principles of the human decision maker, they provide a form of knowledge elicitation suitable for look-ahead systems in a variety of domains; they provide a basis for evaluating prototype and partial systems; they provide manual alternatives in the interim, which self-adapt, create new insights, and continually produce their own methodological enhancements for use in system design and concurrent manual decision tasks
Keywords :
decision support systems; decision theory; forecasting theory; human factors; user interfaces; coherent management; competitors; concurrent manual decision tasks; control system; forecasting methods; human decision makers; information-decision-action systems; knowledge elicitation; look-ahead systems; manual alternatives; markets; operational decision making; partial systems; physical environment; predictive principles; problem domains; prototype; qualitative methods; quantitative techniques; ragged interface; strategic decision making; system design; technological decision support; uncertain environments;