Optimizing Energy Consumption for Smart Home using Machine Learning Techniques

Optimizing Energy Consumption for Smart Home using Machine Learning

Abstract

In rapidly increasing cities, rising energy demand from smart appliances needs effective energy management. This work uses machine learning (ML) and heuristic-based approaches to optimize energy usage in smart homes (SH) by utilizing renewable and sustainable energy resources (RSER) and energy storage systems (ESS). Various optimization techniques, including as genetic algorithm (GA), binary particle swarm optimization (BPSO), wind driven optimization (WDO), bacterial foraging algorithm (BFA) and genetic modified particle swarm optimization (GmPSO), are used to reduce electricity expenditures, peak-to-average ratio (PAR), and carbon emissions while maintaining user comfort. Three energy optimization scenarios are analyzed: Condition 1, which schedules household appliances without renewable energy, achieves 84.09% carbon emission reduction, 89.23% cost savings, and 68.03% PAR reduction; Condition 2, integrating photovoltaic (PV) systems, shows 99.88% carbon emission reduction, 96.80% cost savings, and 96.57% PAR reduction; and Condition 3, combining solar with ESS, improves load distribution and grid independence, reducing carbon emissions by 20.85%, 19.89% reduction in costs and 90.12% reduction in PAR. These findings illustrate that GmPSO outperform in producing sustainable and cost-effective energy saving solutions, offering useful technique for utility companies, regulators, and SH technology developers.