Comparison of stochastic optimal controls with different level of self-learning

Abstract

In the article the problem of control of a linear dynamic system with a square coefficient of quality was introduced. Depending on knowledge we have about the parameters of system, the principles of control of the system under conditions of full and incomplete information as well as the differences in values of coefficients of quality were presented. Potential usage of the above mentioned principles may pertain to physical systems for which the supreme aim of control is minimization of energetic losses as well as optimal control of a linear system in real-time, and also to computer systems where the pricing of value of information is required.