Forecasting and Strategy Unit is part of the Decision and Administration Systems Laboratory and represents an active part of the School of Electrical and Computer Engineering of National Technical University, providing expertise and training opportunities to future engineers, in a field of rapid technological development.
The Unit consists of faculty members, postdoctoral and doctoral researchers who are involved in the conduct of undergraduate and postgraduate courses in the Department. External researchers of the private sector are also involved in the research field.
Furthermore, the Unit provides the means for high-level research activity to PhDs and postdoctoral researchers, via research projects of National and European character.
In recent years, theses and dissertations have been developed in scientific areas covered by the Laboratory and some more are currently being elaborated. The results of the research activities have also been presented to numerous scientific conferences and authoritative journals.
The Unit participated in the M3 international forecasting competition, which ended in Washington in July 1999, presenting the THETA model, a method developed by the Unit during the previous year. The THETA method dominated the competition, producing much better forecasts than all other methods competed. The method outperformed both the existing fully automated, internationally recognized software packages (Forecast Pro, Autocast, Autobox etc.) and methods based on human judgment and experience, developed by academic groups of universities of international reputation (INSEAD, Imperial College, Wharton School, University of Pennsylvania).
In 2018, the Unit co-organized the M4 – Competition together with the University of Nicosia, the most recent and largest forecasting competition of the M series. The M4 involved submissions from many universities (University of Bath, Lancaster University, Harvard, University of Oxford, Georgia Institute of Technology, Monash University) and companies (Uber, ProLogistica, Forecast Pro, Smart Cube, Oracle, AutoBox) worldwide. Its results revealed innovative forecasting approaches, mainly machine learning ones, indicating new, promising fields of research and applications.
The research field of the Unit covers a wide range of areas in an attempt to reconcile the needs of the Greek industry and organizations of the public sector with the guidance of fundamental and applied research in countries of the European Union.
Finally, significant investments in equipment to support the operation of the Unit have already been made or are still in progress, in an effort to maintain the technostructural forefront of research means and laboratory practice in the areas concerned.