An approximate reasoning model for natural gas pricing

The tactical pricing of natural gas in the current industry environment is a complicated and uncertain proposition. Existing modeling systems do not explicitly consider the underlying uncertainty of modeling parameters and thus provide no indication to the user of the uncertainty inherent in the modeling results.; The research described in this dissertation develops a Decision Support System (DSS) for the short-term pricing of natural gas. The uncertainty of system input parameters is explicitly modeled using a non-traditional paradigm known as possibility distributions. Techniques for the application of this paradigm in a computer-based modeling system are developed including a specialized combination technique for the blending of estimates from multiple experts. The DSS implementation includes rule-based pricing and production control, and both textual and graphical output. The application of the model is illustrated using a sample case study.