Abstract

The optimization model generally assumes complete data is known. But in reality, a lot of data is not known for certain. This uncertainty problem can be solved by several approaches, one of which is robust optimization. Uncertainty parameters in optimizing robust are solved by using the set of uncertainties. However, the set of uncertainties yields less conservative results to be applied to the data as a whole. With the abundance of data in recent years, the determination of the set of uncertainties can be done based on data, this method is called Data-Driven Robust Optimization (DDRO). Robust optimization based on data with a machine learning approach presents new challenges. This paper reviews several papers on DDRO and their applications on inventory, scheduling, portfolio selection, industries, and transportation issues.