Article ID: | iaor1996526 |
Country: | Netherlands |
Volume: | 59 |
Issue: | 1 |
Start Page Number: | 102 |
End Page Number: | 122 |
Publication Date: | May 1992 |
Journal: | European Journal of Operational Research |
Authors: | Morecroft John D.W., van der Heijden Kees A.J.M. |
Keywords: | artificial intelligence: expert systems, behaviour |
A group of senior managers and planners from a major oil company met to discuss the changing structure of the oil industry and with the purpose of improving group understanding of oil market behaviour for use in global scenarios. This broad ranging discussion led to a system dynamics simulation model of the oil producers. The model produced new insights into the power and stability of OPEC (the major oil producers’ organization), the dynamics of oil prices, and the investment opportunities of non-OPEC producers. The paper traces the model development process, starting from group discussions and leading to working simulation models. Particular attention is paid to the methods used to capture team knowledge and to ensure that the computer models reflected opinions and ideas from the meetings. The paper describes how flip-chart diagrams were used to collect ideas about the logic of the principal producers’ production decisions. A sub-group of the project team developed and tested an algebraic model. The paper shows partial model simulations used to build confidence and a sense of ownership in the algebraic formulations. Further simulations show how the full model can stimulate thinking about producers’ behaviour and oil prices. The paper concludes with comments on the model building process.