Real-time multiobjective optimization for home load scheduling on an embedded device using an integrated SCADA–AI system

Abstract

Home energy management systems (HEMS) are designed to monitor and optimize residential energy consumption. This paper proposes an approach that integrates artificial intelligence (AI)-based optimization with a supervisory control and data acquisition (SCADA) system within an HEMS implemented on a Raspberry Pi embedded device using Node-RED and Python. The novelty of this work lies in enabling real-time SCADA monitoring and AI-based optimization on a single embedded platform. The theoretical implication of this study is the conceptualization of an integrated edge-based optimization design for HEMS, where multiobjective scheduling is embedded within the operational layer of the system architecture, contributing to the development of decentralized and edge-enabled energy management systems. The contributions include: (1) the development of an integrated SCADA–AI framework for real-time load scheduling on an embedded device; (2) the formulation of a multiobjective optimization model that minimizes grid-imported energy, maximizes photovoltaic (PV) utilization, and reduces the peak-to-average ratio (PAR); and (3) the comparative evaluation of three meta-heuristic algorithms—genetic algorithm (GA), particle swarm optimization (PSO), and ant colony optimization (ACO)—using real IoT-based SCADA monitoring data. The system schedules household loads, including the water heater, rice cooker, washing machine, and microwave oven. Experimental results show that GA achieved the best performance, reducing grid energy by 6.91%, increasing PV utilization by 14.91%, and lowering PAR by 24.44% compared to random scheduling. The execution times of GA, PSO, and ACO on Raspberry Pi 5 were 0.70 s, 0.38 s, and 1.42 s, respectively, demonstrating suitability for real-time embedded applications.