تلفيق همبافت و يادگيري عاطفي تفاوت زماني در مهندسي كنترل

COMBINING CONTEXT AND EMOTIONAL TEMPORAL DIFFERENCE LEARNING IN CONTROL ENGINEERING

One of the most important problems of delayed and nonlinear systems is to fulfill multiple goals simultaneously and in the best conditions. This paper present a method for controlling systems with multiple goals. The method is based on context and has a neuro-fuzzy structure with capability of temporal difference learning. The porposed method, regarding the current status, prior system conditions, and current control goals, would be capable of controlling the system in a way that these golas are achieved in the best way and the least time. In order to clarify the issue and prove the capabilities of the proposed method two well-known control problems, achieving the multiple goals of which, in the best way and the least time, would be very difficult through manipulating other control methods, would be faced using the proposed method.