Posted in | News | Quantum Computing

Quantum-Enhanced Model Predictive Control for Optimized Building Energy Management

A recent study published in Engineering introduces a groundbreaking method for building energy management that integrates quantum computing with model predictive control (MPC), with the goal of improving energy efficiency and advancing building decarbonization efforts.

Buildings are among the largest energy consumers worldwide, playing a substantial role in global energy demand and greenhouse gas emissions. In response to this challenge, researchers Akshay Ajagekar and Fengqi You from Cornell University have developed an adaptive model predictive control (MPC) strategy based on quantum approximate optimization.

This strategy is tailored for buildings that incorporate battery energy storage and renewable energy generation systems, including photovoltaic (PV) panels.

At the core of this strategy is a learning-based parameter transfer scheme for the Quantum Approximate Optimization Algorithm (QAOA), which utilizes Bayesian optimization and Gaussian processes to predict initial quantum circuit parameters. This significantly reduces the computational load of QAOA while allowing the system to adapt dynamically to changing building conditions and external disturbances.

By formulating the MPC problem as a Quadratic Unconstrained Binary Optimization (QUBO) problem, the method enables the efficient computation of optimal control actions to minimize a building’s net energy consumption.

The researchers validated their approach through computational experiments using data from two buildings on Cornell University's campus. They compared their quantum computing-based MPC strategy with both deterministic MPC and quantum annealing methods.

The results demonstrated substantial gains. The quantum MPC strategy delivered a 6.8% increase in energy efficiency over deterministic MPC and achieved a notable 41.2% annual reduction in carbon emissions by efficiently managing battery energy storage and renewable energy sources.

Moreover, the proposed strategy exhibited strong adaptability, effectively adjusting heating and cooling loads in response to ambient temperature fluctuations to maintain indoor comfort while optimizing energy consumption. In terms of computational efficiency, the learning-based QAOA initially required more iterations during the exploration phase; however, as the system evolved, the number of required iterations dropped significantly, ultimately surpassing quantum annealing in performance.

However, the study also noted certain limitations. The building energy system model employed was relatively simplified, and applying the approach to more complex systems could pose challenges due to the increased number of variables, potentially stretching QAOA’s current capabilities. Additionally, although the learning-based method implicitly addresses uncertainties, the integration of explicit uncertainty quantification techniques could further improve the system’s robustness and reliability.

Despite these challenges, the research points to a promising future for building energy management. Enhancing the approach by incorporating real-time carbon intensity data, validating it across a wider range of building types, expanding it to handle more complex control scenarios, and refining the underlying quantum algorithms could further boost both its effectiveness and real-world applicability.

Journal Reference:

Ajagekar, A., & You, F. (2025). Decarbonization of Building Operations with Adaptive Quantum Computing-Based Model Predictive Control. Engineering. doi.org/10.1016/j.eng.2025.02.002.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.