Historic data can be useful to make predictions and thus is an important ingredient in decision support. Which algorithms and models that can benefit from a large amount of data, and provide the optimal balance between computation time and solution quality is an important challenge for our research group.
Sometimes the raw data can be seen as scenarios and be directly used in supporting planning/scheduling decisions. However, often also knowledge about the current state is available, and predictions can be improved by combining this with the historic data (see e.g. our publications on Planning under Uncertainty with Weighted State Scenarios, and Intention-Aware Routing of Electric Vehicles). Furthermore, the latter paper includes an example where a predictive model is built from these two sources of information that facilitates the (route) planning problem.