The challenge of any corporation or government is to maximize the creation of some (commercial or social) value. The only way to create value is by making decisions. The decisions target an uncertain future and are almost always based on incomplete information. Proper methods for decision support are therefore essential. New sensor technologies are emerging within many areas, and an increasing amount of data is now available to industrial companies and societal authorities, but without proper methods for decision support, they are of little value. Recent trends within artificial intelligence and machine learning suggest that one could base decisions on observed data alone, that is, without reference to an underlying theory. When such a theory is available, however, sound decision support should take mutually into account theoretical process understanding, observed data, uncertainties in the process model and uncertainties in the data. This can be achieved by combining computer simulations with the ensemble-based methodology for data assimilation, uncertainty prediction, optimization, and decision making.
The Data Assimilation and Optimization group at NORCE Energy has world-leading expertise on ensemble-based methods for decision support. The main application in the last two decades has been petroleum reservoir management, where the group currently hosts two major projects, Digires, and 4DSEIS, supported by 10 petroleum companies and the Research Council of Norway. In recent years, the group has also been actively engaged in other applications, such as CO2 sequestration,geothermal energy, hydropower, and medicine.