Prediction and optimization in a dynamic environment: a case study

This application-oriented paper presents a particular case study: a general description of the design and implementation of an intelligent system for one of the world's largest car manufacturers. The presented optimization problem was very challenging, as it included many prediction and optimization elements. The system was installed as a full production system in 2002, and has been used since then to recommend the best distribution of cars on a daily basis. The system improves profits by hundreds of millions of dollars (USD) per year. This significant return on investment was possible only by combining prediction, optimization, and adaptation techniques into one integrated system