Multiple Scenario-based Model Predictive Control with Decision Time Limit Determination of Scenario Selection

Abstract

In the Cyber Physical System or the Digital Twin, the control strategy generates multiple scenarios in the cyber world with optimization of future models and objective functions, and these scenarios are utilized to determine an optimal strategy, which is then applied to the physical world. In these procedures, the decision-making to select and fix a future scenario and its time limit are important factors. In this study, considering the scenario decision time limit, a procrastination strategy is introduced and formulated as a new model predictive control framework. It is to postpone the decision and preserve the freedom of scenario choice for the future. In the proposed method, the concept of a common admissible set for control trajectory and its branch point are introduced. A simple numerical example and an application to an energy management problem are shown to illustrate and verify the effectiveness of the proposed method.