Polynomial Load Curve Modeling and Coyote Optimization Algorithm for Demand Side Management Under Time-of-Use Tariffs

The paper’s main contribution is applying the Coyote Optimization Algorithm (COA) to non-domestic TOU load optimization — a gap in literature, which has mostly focused on residential demand response with conventional algorithms (PSO, GA, GWO). Using real half-hourly campus data (UiTM Permatang Pauh), it develops sixth-degree polynomial load-profile models (weekday/weekend, R² up to 0.98) under Malaysia’s RP4 tariff, then uses COA to shift load from peak to off-peak periods without altering total energy or maximum demand. Results show modest but real cost savings (RM856.45, 0.18% in November; RM778.68, 0.19% in December), validating COA’s stability and fast convergence for institutional DSM.