Probabilistic production costing

Abstract New production cost algorithms are compared with representative algorithms used by the electric utility industry. Accuracy and efficiency of new and conventional algorithms are assessed using the framework of probabilistic production costing. Their usefulness in the context of long-term capacity expansion models is evaluated. Accuracy is considered in two contexts: first, in fitting original customer demand; and second, in approximating generator forced outages. Comparisons of algorithm efficiency are based on an EPRI synthetic system with dispatch of 174 units. Conventional algorithms are drawn from models in use by TVA, Southern Company and MIT. The new algorithms are based upon continuous functions for fitting the equivalent load curve.