Multi-objective optimization using genetic algoritm for efficient energy management in smart homes

Abstract

The ever-growing load demand and irregularity in the electricity load profile, especially in residential areas, have led to a surge in electricity prices. Rapid advancements in the electricity market and Renewable Energy Systems (RES) have spurred extensive research into energy management through demand side management (DSM) expedited by Smart Home Energy Management Systems (SHEMS). In countries such as Taiwan, where Real-Time Pricing (RTP) tariff schemes are used, efficient energy management can be achieved by utilizing optimization algorithms. The focus of this study was to use Genetic Algorithm (GA), a nature inspired optimization algorithm, to achieve efficient energy management in smart homes via Multi-Objective Optimization (MOO). Three objectives are optimized for the home user: namely electricity cost, user comfort, and peak-to-average ratio (PAR). The scheduling problem not only aims for maximum user satisfaction but also considers two user interruption parameters: with penalty and without penalty. The results have shown a 14.56% cost reduction in scheduling without user interruption, 18.62% cost reduction in scheduling considering user interruption (with penalty), and 15.69% cost reduction in scheduling considering user interruption (without penalty). The maximum user comfort was improved by 67.48% (without user interruption), 62.62% (user interruption with penalty) and 41.65% (user interruption without penalty), and the PAR was reduced by up to 51.53% on average. Despite the stochastic nature of electricity consumers, with an optimization system, the cost and peak demand can be curtailed significantly while still maximizing their comfort level.