Abstract
The concept of goal uncertainty in a generalized information system is modeled by treating utility functions as interval valued rather than real valued. Furthermore, since the system is dynamic, the intervals shrink as information accumulates. The lower (upper) endpoints of these intervals are shown to be sub(super) martingales. Decision rules are formulated to select the appropriate course of action for each cycle through the decision-information feedback loop which is the heart of this information system. Finally, the theory of multifunctions is used to obtain convergence theorems for the estimated expected utilities generated by this system
∗This research was supported in part by National Science Foundation Grant IST-83-03900
∗This research was supported in part by National Science Foundation Grant IST-83-03900
Notes
∗This research was supported in part by National Science Foundation Grant IST-83-03900