Control of supply networks by robust optimal control and using observers

The control strategy to organize the flow of goods is extremely important in managing supply networks. Handling of uncertainty, e.g. customer demand, is often a considerable challenge. We introduce methods of constrained robust optimal control, a technique from control theory, to compute the explicit control strategy for small-sized demand-driven discrete-time controlled dynamical systems with uncertainties. This allows us to avoid any assumptions about the form of the policy a priori. We aim to show the applicability of the methods, instead of finding innovative policies as the resulting policies are well known. Another control problem arises when control actions depend on the information of multiple states and these states are not (completely) known, e.g. due to communication issues. We use observers, another technique from the control theory, to derive the required state information based on the in- and output. By means of an example, it is shown that by using these observers, the control policy can be perfectly utilized and no longer depends on communication.