LLM agent optimizes building energy system costs efficiently
AutoEnergyDesign: An LLM-Driven Agent for Automatic Energy System Design Optimization
Artificial Intelligence
Summary
Designing energy systems for buildings is very complex and usually takes a lot of time because it involves testing many options over a full year in simulation. The authors created AutoEnergyDesign, a smart computer agent that uses large language models to learn from past simulation results and suggest better designs automatically. It keeps track of what works and what doesn’t, making the search much faster. Their tests showed the agent finds near-best energy system designs with far fewer simulations compared to traditional methods.
What this means in practice
- •For building design teams: Automatically optimize long-term energy system configurations with fewer simulations and reliable cost outcomes.
- •For energy management software developers: Integrate LLM-driven agents to improve optimization speed and decision tracking in building energy design tools.
Authors
Amedeo Buonanno, Salvatore Fabozzi, Maria Valenti
Abstract
Designing building energy systems is a high-dimensional, mixed-integer search problem: when each design is evaluated by a year-long simulation, black-box optimizers typically need hundreds of evaluations. This paper presents AutoEnergyDesign, a framework in which a Large Language Model (LLM) agent reasons over previous simulation results, proposes configuration changes and learns in context from each outcome, while a git-based mechanism automatically reverts unsuccessful changes. The agent couples a reasoning model with an explicit memory organisation (procedural, semantic, episodic and reflective), a constrained action interface and an environment-grounded keep/revert signal. The simulator covers eight technologies (photovoltaics, micro-wind, combined heat and power, solar thermal, gas boilers, batteries, heat pumps and inverters) and computes the 25-year Net Present Cost (NPC), including capital and operating costs and salvage value. On a case study with measured building load profiles, the agent, run with Claude Haiku 5.5, Sonnet 5.5 and Opus 5.5 (ten runs of 50 simulations each), is compared with random search, Tree-structured Parzen Estimator (TPE) Bayesian optimization and a $(1+1)$ evolution strategy sharing the agent's acceptance rule (300 evaluations, ten seeds). With all three models the agent reaches an NPC of about 3.53 MEUR, 1.20% below the best TPE result and 0.23% below the best evolution-strategy result. It comes within 1% of this design after a median of 18 (Opus) to 27.5 (Haiku) simulations, whereas among the baselines only the evolution strategy does so, in half of its seeds and after a median of 192 evaluations. Opus reaches the best design in all ten runs, Sonnet in six (two runs never come within 1%), and Haiku comes within 1% in every run but reaches it exactly only once. Every decision is recorded in an auditable trace.