Two-level hierarchical optimal control based on interaction principle for large scale systems

In this paper a new two-level hierarchical control is proposed for optimal control of large scale discrete linear systems. The proposed method significantly decreases computational complexity of the solution in comparison with the centralized solution. On the other hand it results in a slight deviation in global optimal results. However, the gain of complexity that has been obtained is more important and considerable. In centralized solution, it is required to calculate the inverse of large-size sparse matrices which are very complex to compute or do not exists because of zero determinants. However the proposed solution results in non-sparse matrices of very lower order. The proposed method is simulated and the obtained results are compared with the centralized optimal control.