Model-Assisted Optimal Control for Split Air Conditioners (ACs)

Learn about innovative energy-saving solutions for cooling systems in this lecture by Prof. Vinayak Naik from BITS Pilani, Goa. The talk presents a Model-assisted Optimal Control (MaOC) algorithm designed to optimize energy consumption of split air conditioners, which account for over 50% of buildings' energy usage. Discover how this algorithm observes thermal environments, measures AC efficiency, and generates optimal execution schedules using convex function properties and MAXMIN problem formulation. The presentation covers real-world evaluation results showing 83% energy savings compared to greedy techniques and 34% faster cooling times, with performance comparable to more complex Reinforcement Learning approaches. Prof. Naik also discusses anomaly detection for AC maintenance and shares insights from this award-winning research conducted with PhD student Keshav Kaushik, which received the best CS PhD dissertation award at BITS Pilani, Goa for 2024.