Forecasting Short Term Power Prices in the Ontario Electricity Market (OEM) with a Fuzzy Logic Based Inference System

The OEM, which opened in May 2002, is relatively new and is still under change. In addition, the bidding strategies of the participants are such that the relationships between price and fundamentals are non-linear and dynamic. Therefore, a flexible model is needed to achieve good forecasting in OEM. This paper uses a Takagi-Sugeno-Kang (TSK) fuzzy inference system in forecasting the one-day-ahead real time peak price of the OEM. The forecasting results of TSK are compared with those obtained by statistical and neural network based forecasting. The comparison suggests that TSK has considerable value in forecasting one-day-ahead peak price in OEM