A time series approach to short term load forecasting through evolutionary programming structures

Abstract

Multiple local minimum points often exist on the surface of forecasting error function of the time series models. Solutions of the traditional gradient search based identification technique, therefore, may stall at the local optimal points which lead to an inadequate model. By simulating natural evolutionary process, the evolutionary programming (EP) algorithm offers the capability of converging towards the global extremum of a complex error surface. The EP based load forecasting algorithm is developed to identify the autoregression moving average (ARMA) model for one week ahead hourly load demand forecasts. Numerical tests indicate the proposed EP approach provides a method to simultaneously estimate the appropriate order and parameter values of the ARMA model for diverse types of load data. Comparisons of forecasting errors are made to the traditional identification techniques used by SAS statistical commercial package.