Demand elasticity analysis by least squares support vector machine

Demand-side management (DSM) has been introduced and paid much effort in the electricity industry for such a long time, and become more and more focused in the most of the electricity markets recent years, because of its effect on the modification of consumer demand for energy through various methods especially financial incentives. But due to the complexity of the influence factors, the demand elasticity analysis on electric load is along one of focused and unsolved problems in the researches of electricity market. This paper presents a mathematical tool based on optimization models that is intended to be used to evaluate the load elasticity. Here one effective way of tackling the elasticity computation problem is proposed to simulate the accurate demand response for its influence factors, where least square support vector machine (LS-SVM) is generalized to build the simulation model. First to overcome the limitations of traditional parametric regression methods based on least square rule, load mapping model are accomplished by LS-SVM that is a strong nonlinear regression tool and outperforms standard support vector machine in the regression accuracy and modeling velocity by obtaining the solutions of a set of linear equations instead of solving a quadratic programming problem. Based on that model, numerical simulation is performed to simulate the changes in load and its influence factor in consistency with the latent mapping relationship. Then demand elasticity models are deduced, whose elasticity coefficients are computed by the generated samples. The method embodies clear concepts, so it can be easily understood theoretical and computationally realized. Case studies are used to test the proposed model. In this scenario, the goal is to show the response of demands to different influence signals. Alternative analyses based on assessing the elasticity of power to other input parameters could be also carried out.