Planning of today's electric utilities demand careful consideration of issues such as environment, demand-side management, non-utility generation, and new technologies which are subject to different constraints and uncertainties. Utilities have long developed and used models for their short and long-term planning, most of which are single purpose, large, data intensive, and do not fully account for uncertainties.

New techniques have emerged to deal with uncertainties in utility planning. Among them, the Analytic Hierarchy Process (AHP) has been more successful in assessing uncertainties, and found to be well structured and applicable to individual as well as group decision makers. However, the results of this method are merely point estimate values.

It is the objective of this research to identify a methodology which is capable of evaluating uncertainties with relative ease and accuracy without the need for a large volume of data and complicated software packages. The Analytic Hierarchy Process has been extended to estimate the variance of the error in judgments and therefore the confidence interval of values instead of point estimate values. A simulation study was carried out to check the accuracy of error variance (QI) in confidence interval calculations. The results showed that QI has a linear relationships with the variance of weights.

The extended AHP method is applied to three case studies, including 1) Third party generation bidding evaluation criteria, 2) Identification and evaluation of different load management programs on utility peak reduction, and 3) Oil price prediction for electric utilities. This method promises to be an effective decision making tool for evaluating uncertainties in electric power system planning.