Control-theoretic results on dynamic decision making

Several approaches to learning in dynamic decision making tasks are developed in this paper on the basis of the application of feedback control theory to the case study of the Sugar Production Factory task. Previous experimental models are not robust to workload change and require a large amount of information to be stored. The control model presented here not only avoids such shortcomings, but also significantly enhances the system efficiency, adaptivity and robustness. On the other hand, as trust and self-confidence are closely linked to the capacity of automation and manual control in a supervisory control system, it behooves us to develop a dynamic model to assist the operator in gaining a better understanding of capacities. A quantitative model of trust in automation is then proposed to accurately characterize operator's reliance on automation. Those results are demonstrated through simulation within a framework of a Sugar Factory supervisory control system.