DECISION-MAKING IN CONDITIONS OF EXTREME UNCERTAINTY.

In the last few years management has been forced to take decisions in conditions of extreme uncertainty. One consequence of this has been an increase in the use of formal methods of analysis in major decision problems. This paper describes some observations on how companies were handling decision-making under uncertainty, made during a S.S.R.C. sponsored research project on the influence of computer methods in planning and decision-making, viz. 1. Managers were consciously handling uncertainty by evaluating their major decisions in the light of several possible alternative forecasts. 2. Statistical decision theory was not used because, though in theory it was ideally suited for this situation, in practice management had the following objections to the use of this technique: (a) It assumed an accuracy of subjective probability estimates which was unrealistic in real life. (b) It required the use of artificial criteria of choice which were totally unacceptable to senior management. (c) It over-simplified the decision-making process. 3. In the thirty-two companies visited, management was using a systems analysis approach to decision-making involving the use of decision trees to structure the decision. Thereafter, instead of carrying out a probalistic analysis, a full evaluation to determine the quantitative and qualitative results would be carried out for each feasible path though the tree. This would be followed by an iterative elimination process. 4. Management were also endeavouring to cope with uncertainty by being more adaptive and flexible in their decision-making and are developing contingency or 'fall-back' strategies. Plans were not looked on as rigid blue-prints, but more as loosely linked frameworks of decision trees with the actual path through the decision tree dependent on future events. [ABSTRACT FROM AUTHOR]