Load Scheduling and Effective Price Prediction Utilizing Multi-Armed Bandit-based Multi-Objective Optimization Approach

This paper represents a novel strategy for the internet of things (IoT)-based smart home energy management (HEM) to control energy use and maintain customer loyalty effectively. Intelligent scheduling of home appliances from peak to off-peak hours is essential for lowering the power consumption cost, while preserving customer comfort. Hence, the scheduling management based on day-ahead price (DAP) and critical peak price (CPP) tariff scheme is discussed for effectively minimizing the power consumption cost, peak-to-average ratio (PAR), and maximizing the profit of the consumer. The multi-objective optimization problem is handled by the authors’ proposed multi-objective American zebra optimization algorithm (MOAZOA) for obtaining the ideal scheduling pattern. Additionally, two distinct scenarios for load scheduling are considered: one without renewable energy system (RES) and another with RES. Further, an advanced price prediction algorithm based on a hybrid multiarmed bandit (MAB) with MOAZOA approach is proposed for minimizing the electricity consumption cost, while simultaneously maximizing the profit of the supplier. The efficacy of the proposed algorithm is compared with the multi-objective Arithmetic optimization algorithm (MOAOA), multi-objective Grey wolf optimizer algorithm (MOGWO), and modified multi-objective particle swarm optimization (MMOPSO). It is observed that in comparison to unscheduled load with real-time price tariff scheme, the total cost of the day is reduced to 54%, 65%, 25%, and 74% by employing MMOPSO, MOGWO, MOAOA, and MOAZOA, respectively, without even incorporating the RES system. Furthermore, the performances of these algorithms are compared considering various statistical data.

Cite this article as: D. Das, R. K. Khadanga and D. K. Rout, "Load scheduling and effective price prediction utilizing multi-armed bandit-based multi-objective optimization approach," Electrica, 2026, 26, 0188, doi: 10.5152/electrica.2026.25188.