An ordinal optimization based method for selecting events

Event-based optimization (EBO) is a general method for solving optimization problems in discrete event dynamic systems (DEDS). The performance of event-based optimization is closely related to the selection of events, which can be formulated into an optimization problem. The proper ways to select events can be obtained by solving this problem. However, the solution space is huge and the structure is usually unknown. The common way to select events is based on experience. It remains open how to select events in large scale DEDS without prior knowledge. We consider this important problem in this paper and make two major contributions. First, we formulate the event selection problem in EBO into an optimization problem. Second, we propose an ordinal optimization based method to select events without prior knowledge. The performance of this method is demonstrated by a toy example for proof of concept and an electric vehicle (EV) charging control problem. Numerical results show that the proposed method works well in selecting events.