Tarx Models for Spikes and Antispikes in Electricity Markets

Abstract

Two electricity price models are presented in the frame of switching and threshold autoregressive exogenous (TARX) models, both in continuous time and in discrete time. The first model is based on the biologically inspired McKean model for spiking neurons, the second model is its extension crafted in such a way to easily display a spike and antispike phenomenology, showing random spikes only in price crest phases (daytime) and random antispikes only during price trough phases. The spiking behavior can be explained by the presence of a mathematical threshold mechanism that can be related to power grid congestions. Revised and published with the same title in Energy Market (EEM), 2010 7th International Conference on the European.

Keywords: stochastic processes, time series analysis, power system economics