GPU-Accelerated Optimization Solver for Unit Commitment in Large-Scale Power Grids

2025-12-07

AI Summary

To address the computational bottlenecks—excessive runtime and slow convergence—of large-scale power system unit commitment (UC) problems on conventional CPUs, this paper proposes a GPU-accelerated mixed-integer linear programming (MILP) solver. The method leverages GPU parallelism to implement the Primal-Dual Hybrid Gradient (PDHG) algorithm for efficiently solving linear relaxation subproblems, thereby accelerating bound estimation and branch-and-bound search. It further incorporates customized sparse matrix handling and memory optimization techniques to enhance scalability in high-dimensional, long-horizon scheduling scenarios. Evaluated on real-world power systems with 4,224–6,717 buses, the solver achieves speedups of several-fold to over an order of magnitude compared to state-of-the-art CPU-based solvers, while rigorously preserving solution feasibility and optimality. This advancement enables efficient, reliable, and near-real-time rolling optimization for modern power system dispatch.

Technology Category

• Search and Optimization: Mixed Discrete/Continuous Search
• Constraint Satisfaction and Optimization: Mixed Discrete/Continuous Optimization
• Planning, Routing, and Scheduling: Mixed Discrete/Continuous Planning

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

This work presents a GPU-accelerated solver for the unit commitment (UC) problem in large-scale power grids. The solver uses the Primal-Dual Hybrid Gradient (PDHG) algorithm to efficiently solve the relaxed linear subproblem, achieving faster bound estimation and improved crossover and branch-and-bound convergence compared to conventional CPU-based methods. These improvements significantly reduce the total computation time for the mixed-integer linear UC problem. The proposed approach is validated on large-scale systems, including 4224-, 6049-, and 6717-bus networks with long control horizons and computationally intensive problems, demonstrating substantial speed-ups while maintaining solution quality.

Problem

• Accelerates unit commitment solving via GPU-accelerated PDHG algorithm
• Reduces computation time for large-scale mixed-integer linear UC problems
• Validates speed-ups on large power grids while preserving solution quality

Innovation

• GPU-accelerated solver for unit commitment problem
• Uses Primal-Dual Hybrid Gradient algorithm for subproblems
• Validated on large-scale power grids with speed-ups