Relations between subjective uncertainty and performance in complex problem solving (based on the management of virtual factory)

Abstract

The process of complex problem solving while exploring and managing a virtual factory simulated by a computer dynamic scenario is considered. The relations between: (a) estimates of subjective uncertainty (beliefs about features of the system explored and managed, methods of the problem solving, expectations of effects of one’s actions), and (b) performance in complex problem solving (measured as increase/decrease of amount of virtual money) are identified. It is shown that subjective uncertainty has a complex structure and affects complex problem solving in different ways. It can be a powerful factor of positive motivation (via interest in the task) or a factor of negative motivation and desire to escape from the complex problem solving situation. There were two different mechanisms of uncertainty reduction in different types of the problem solvers. The solvers with exploratory behavior demonstrate step-by-step increase of optimistic certainty and confidence based on more and more successful orientation in the system and methods of work with it. The solvers with chaotic or rigid behavior demonstrate increase of pessimistic certainty caused by repeating failures, growing confidence in their inability to cope with the task, and desire to get out of the complex problem situation. A regression model of relations between subjective uncertainty and efficiency of complex problem solving is proposed.