Robust Model Predictive Control of Constrained Supply Networks via Invariant Ellipsoids Technique

Abstract

The problem of robust control strategy synthesis for distributed supply network under demand uncertainty, time delays and state and control constraints is considered. An invariant ellipsoids approach is used for robust control problem solving, since the uncertain demands are regarded as an external disturbance. On the base Model Predictive Control approach, the designed control law implements in the form of linear feedback signal based on mismatch between the current state and safety stock level and provides external disturbances effect suppression with simultaneous robust stabilization of closed-loop system. Via invariant ellipsoids technique the considered problem was presented in the terms of Linear Matrix Inequalities and a solution of corresponding semi-definite optimization problem was also obtained. As an example, the three-tier supply network with five nodes robust control problem is considered.