Large Language Models for Solving Economic Dispatch Problem
Sina Mohammadi, Ali Hassan, Rouzbeh Haghighi, Van-Hai Bui, Wencong Su
TL;DR
The economic dispatch (ED) problem aims to minimize generation cost while satisfying power balance and generator limits, formally min ∑_{i=1}^{N} (a_i P_{G_i}^2 + b_i P_{G_i} + c_i) subject to ∑_{i=1}^{N} P_{G_i} = P_D and $P_{G_i}^{\min} ≤ P_{G_i} ≤ P_{G_i}^{\max}$ for all i. The authors investigate off-the-shelf large language models (LLMs) using few-shot prompting to solve ED in a single interaction, circumventing explicit mathematical prompts. They compare non-evolutionary prompting and evolutionary prompting (LLM-guided crossover/mutation) on the IEEE 118-bus system with 19 generation units, using PGLib-OPF data and a Gurobi baseline for reference. Results show that LLMs can produce plausible dispatch vectors and near-optimal costs, with non-evolutionary prompting achieving better constraint satisfaction and evolutionary prompting offering a semi-guided search; together, the approach suggests a practical, data-light alternative for ED when foundational grid models are available. The work points to extending these ideas to more complex ED variants, including DC/AC OPF, and broader loading scenarios.
Abstract
This paper investigates the capability of off-the-shelf large language models (LLMs) to solve the economic dispatch (ED) problem. ED is a hard-constrained optimization problem solved on a day-ahead timescale by grid operators to minimize electricity generation costs while accounting for physical and engineering constraints. Numerous approaches have been proposed, but these typically require either mathematical formulations, face convergence issues, or depend on extensive labeled data and training time. This work implements LLMs enhanced with reasoning capabilities to address the classic lossless ED problem. The proposed approach avoids the need for explicit mathematical formulations, does not suffer from convergence challenges, and requires neither labeled data nor extensive training. A few-shot learning technique is utilized in two different prompting contexts. The IEEE 118-bus system with 19 generation units serves as the evaluation benchmark. Results demonstrate that various prompting strategies enable LLMs to effectively solve the ED problem, offering a convenient and efficient alternative. Consequently, this approach presents a promising future solution for ED tasks, particularly when foundational power system models are available.