A Stochastic Dynamic Programming model for scheduling of offshore petroleum fields with resource uncertainty

Norwegian deliveries of natural gas to Europe have grown considerably over the last years. Additionally, plans involve even greater supplies, introducing major gas fields such as the Troll field. The market for natural gas may to a large extent be viewed as a contractual market. This is normally explained by the big investments involved in development and transport. In such a perspective, the supplier's planning problem of scheduling fields and pipes may prove important in order to be able to meet contractual agreements. This paper describes a model of Stochastic Dynamic Programming type which may be viewed as a first attempt in solving this type of problem. The main focus in this model is on project scheduling and resource uncertainty. Each project's production profile is viewed as a stochastic variable. Then a possible goal could be to minimize expected deviation from a given predicted contract profile. We use the term SPSP (Stochastic Project Scheduling Problem) to refer to our problem. Besides a simplified description of the mathematical model, the paper also describes some tests with model-examples. These examples are constructed to obtain some interesting (non-intuitive) effects. The paper concludes with some remarks on complexity and speedup possibilities.