A Novel Flower Pollination Method for Unit Price Estimation in a Microgrid

A microgrid requires scheduled power dispatch and proper forecasting of energy for effective management of all the components in the power system and a reduction in operating costs. The operational expenses increase due to the variable source of energy and load demand. Electricity price forecasting plays an important role as the traditional grid also requires precise unit price predictions for trading of electricity in the market. While handling the big data sets that are produced every 15 seconds, the traditional machine learning algorithms are time consuming and may create curve overfittings while predicting the unit price in the microgrid. Hence, in this paper, a hybrid approach has been proposed which combines traditional machine learning methods with a flower pollination algorithm for the estimation of unit price in a microgrid. The proposed method involves selection of characteristics, major component analysis, and a composite system for both optimization and regression analysis. A MATLAB based algorithm has been developed to test the hypothesis and to test the efficiency of algorithm.