Fuzzy Control

We identify the domain of intelligent, distributed control as a fertile area for application of DAI techniques. We use a tank level control problcm (see Figure l) to demonstrate how cooperative controllers can achieve quicker and more accurate control over independently acting controllers. The distribution of control efforts at which the sum of the individual cost functions is minimum determines the allocation of control efforts between the cooperative controllers. We compared the performance of cooperative controllers against independent controllers where each controller sets its control output based on current system state without considering the effects of other controllers. Figure 1: Valves on top/bottom of tank regulate inlet/outlet flows to maintain fluid level; a third outlet is the source of disturbance.