Quantum computing paves the way for low-carbon building operations

A new study published in Engineering presents an innovative approach to building energy management that combines quantum computing with model predictive control (MPC), aiming to enhance energy efficiency and drive decarbonization in buildings.

Buildings are major energy consumers, contributing significantly to global energy use and greenhouse gas emissions. To address these issues, researchers Akshay Ajagekar and Fengqi You from Cornell University developed an adaptive quantum approximate optimization-based MPC strategy. This strategy is designed for buildings equipped with battery energy storage and renewable energy generation systems, such as photovoltaic (PV) panels.

The heart of the strategy is a learning-based parameter transfer scheme for the quantum approximate optimization algorithm (QAOA). It leverages Bayesian optimization and Gaussian processes to predict initial quantum circuit parameters. This not only reduces the computational burden of QAOA but also enables the system to adapt to changing building states and external disturbances. By treating the MPC problem as a quadratic unconstrained binary optimization (QUBO) problem, the approach can compute optimal controls to minimize a building’s net energy consumption.

The researchers conducted computational experiments using data from two buildings on Cornell University’s campus. They compared the performance of their quantum computing-based MPC strategy with deterministic MPC and quantum annealing. The results showed remarkable improvements. The quantum MPC strategy achieved a 6.8% improvement in energy efficiency compared to deterministic MPC. It also led to a significant annual reduction of 41.2% in carbon emissions by effectively managing battery energy storage and renewable generation sources.

Moreover, the proposed strategy demonstrated good adaptability. It could adjust the heating and cooling loads in response to ambient temperature changes, maintaining indoor comfort while optimizing energy use. In terms of computational efficiency, although the learning-based QAOA required more iterations in the initial exploration phase, the number of iterations decreased rapidly as the system evolved, outperforming quantum annealing in this aspect.

However, the study also acknowledged some limitations. The building energy system model used was relatively simple, and for more complex systems, the increased number of variables might challenge QAOA’s current capabilities. Additionally, while the learning-based approach implicitly handles uncertainties, incorporating uncertainty quantification methods could further enhance the system’s reliability.

Despite these challenges, this research offers a promising direction for future building energy management. Integrating real-time carbon intensity metrics, validating the approach across diverse buildings, extending it to more complex control scenarios, and optimizing quantum algorithms could further improve its performance and practical applicability.

The paper “Decarbonization of Building Operations with Adaptive Quantum Computing-Based Model Predictive Control,” authored by Akshay Ajagekar, Fengqi You. Full text of the open access paper: https://doi.org/10.1016/j.eng.2025.02.002. For more information about Engineering, visit the website at https://www.sciencedirect.com/journal/engineering.

END