Optimisation algorithms used in home energy management systems: A review

• An efficient management of household energy consumption potentially can save energy that can contribute to mitigate the climate change impact on the environment. • A household energy consumption typically can be controlled by using smart scheduling appliances considering home energy management systems (HEMS). • Critical review on optimization algorithms used in HEMS categorizing into mathematical, heuristic, and machine learning-based HEMS methods. • Adaptive HEMS that can effectively manage complex energy systems, future research directions highlight the necessity of hybrid optimization approaches that combine multiple algorithms. The growing number of photovoltaic (PV) systems deployed in houses and the increase in energy consumption have made home energy management systems (HEMS) indispensable. As climate concern is important for today’s environment, research into developing energy-efficient technologies has gained momentum and it has become one of the major research topics in sustainable energy including building sustainability. Residential energy usage accounts for approximately 30% of total energy consumption nationwide. Therefore, an efficient management of household energy consumption potentially can save energy that can directly contribute to mitigate the climate change impact on the environment. A household energy consumption typically can be controlled by using smart scheduling appliances. The use of optimization methods by HEMS may greatly lower energy consumption and increase energy uses efficiency. The application of Particle Swarm Optimization (PSO) and the Grey Wolf Optimization (GWO) techniques, for example, have shown energy cost reduction up to 20% to 25%. As revealed, the Advanced Deep Reinforcement Learning (DRL) techniques have demonstrated enhanced adaptability to dynamic environments resulting in up to 13% reductions in peak-hour energy demand. This paper offers a critical review on optimization algorithms used in HEMS categorizing into mathematical, heuristic, and machine learning-based HEMS methods. Analysing the contributions and limitations and real-world applicability of these methods and relevant techniques, the review article highlights recent trends and gaps in existing HEMS research. While conventional strategies like linear programming works well for small-scale systems, newer approaches like Machine Learning (ML) and DRL are found to be more accurate and scalable for dynamic, real-time optimization. To create smarter, more adaptive HEMS that can effectively manage complex energy systems, future research directions highlight the necessity of hybrid optimization approaches that combine multiple algorithms, improved real-time processing capabilities, and the integration of Internet of Things (IoT) and edge computing technologies.