Entropy management of Gaussian stochastic systems

Two approaches to the entropy management of Gaussian stochastic system are considered. The first approach is scalar and implements the concept of "growth points". In this case the problem of maximizing (increasing) or minimizing (decreasing) the system entropy is solved. The second approach is the vector management, allowing to ensure effective changing of the entropy of two-dimensional vector, the components of which are randomness and self-organization entropies. For vector control an optimization problem on the conditional extremum is formulated. This problem can be solved using penalty methods. It is shown that the vector management of entropy for a number of cases has advantages compared to the scalar management. Examples of entropy models of real stochastic systems are provided.