Decision support system for scheduling a Flexible Flow System: Incorporation of feature construction

We study the problem of scheduling a Flexible Flow System (FFS) while utilizing machine learning techniques. Specifically, the effect of incorporating Feature Construction in an existing adaptive learning-based decision support system is investigated. The primary purpose of constructing features is to increase the efficiency of obtaining information from available data by using new features in addition to the initial set of features. There is a need to obtain information effectively while scheduling an FFS. The incorporation of feature construction in such a framework is thus beneficial, which is documented in our preliminary results.