Qubit-Efficient Quantum Annealing for Stochastic Unit Commitment

Wei Hong, Wangkun Xu, Fei Teng

Abstract

Stochastic Unit Commitment (SUC) has been proposed to manage the uncertainties driven by the integration of renewable energy sources. When solved by Benders Decomposition (BD), the master problem becomes a binary integer programming which is NP-hard and computationally demanding for classical computational methods. Quantum Annealing (QA), known for efficiently solving Quadratic Unconstrained Binary Optimization (QUBO) problems, presents a potential solution. However, existing quantum algorithms rely on slack variables to handle linear binary inequality constraints, leading to increased qubit consumption and reduced computational efficiency. To solve the problem, this paper introduces the Powell-Hestenes-Rockafellar Augmented Lagrangian Multiplier (PHR-ALM) method to eliminate the need for slack variables so that the qubit consumption becomes independent of the increasing number of benders cuts. To further reduce the qubit overhead, quantum ADMM is applied to break large-scale SUC into smaller blocks and enables a sequential solution. Consequently, the Quantum-based PHR-ADMM (QPHR-ADMM) can significantly reduce qubit requirements and enhancing the applicability of QA in SUC problem. The simulation results demonstrate the feasibility of the proposed QPHR-ADMM algorithm, indicating its superior time efficiency over classical approaches for large scale QUBO problems under the D-Wave QPU showcases.

The Challenge: The "Qubit Explosion" in Power Systems

The Stochastic Unit Commitment (SUC) problem is a cornerstone of modern power system operations. It requires determining the optimal schedule of generating units to meet demand while accounting for the unpredictable nature of renewable energy. Because it involves discrete binary decisions (on/off) across numerous scenarios, SUC is NP-hard, becoming exponentially difficult to solve as grid complexity increases.

Quantum Annealing (QA) is theoretically well-suited for these binary optimization problems. However, current QA approaches face a significant hurdle: the "qubit explosion." To handle inequality constraints—such as ensuring a generator's output stays within its physical limits—standard methods rely on slack variables. These variables consume a massive number of qubits, often exceeding the capacity of contemporary quantum hardware. Furthermore, in iterative techniques like Benders Decomposition, the number of constraints (and thus required qubits) grows with every iteration, making long-term optimization on current QPUs impossible.

Methods: QPHR-ADMM Framework

To overcome these hardware limitations, the researchers developed a two-pronged strategy to achieve extreme "qubit efficiency":

  1. PHR-ALM (Powell-Hestenes-Rockafellar Augmented Lagrangian Multiplier): Instead of using binary-encoded slack variables to handle inequalities, the authors employ an analytical Lagrangian formulation. This mathematical approach incorporates constraints directly into the objective function, eliminating the need for extra qubits. This ensures the qubit count remains independent of the number of Benders cuts.
  2. D-ADMM (Direct-extended Alternating Direction Method of Multipliers): The framework decomposes the massive SUC master problem into smaller, manageable "blocks" (e.g., grouping variables by individual power plants). These blocks are solved sequentially on the Quantum Processing Unit (QPU).
The QPHR-ADMM architecture, showing the decoupling of the master problem for sequential QPU processing and scenario-based CPU sub-problems.

By combining these methods, the QPHR-ADMM framework allows the master problem to be handled by the QPU while continuous sub-problems (economic dispatch) are processed in parallel by classical CPUs.

Key Insights and Results

The algorithm was evaluated using both a synthetic 4-generator system and the realistic IEEE 118-bus system (featuring 54 thermal units). The results highlight several critical advantages of the QPHR-ADMM approach:

  • Breaking Linear Scaling: In traditional QA, the number of qubits grows linearly with iterations and constraints. QPHR-ADMM keeps the qubit requirement constant, regardless of how many Benders cuts are added.
  • Massive Qubit Savings: In the 4-generator test case, a "Basic QA" approach would have required over 2,106 qubits—exceeding stable embedding limits. QPHR-ADMM solved the same problem using only 24 qubits.
  • Hardware-Aware Scalability: The authors successfully solved the IEEE 118-bus system (1,296 binary variables) by decomposing it into blocks of 30–32 variables, the "sweet spot" for the D-Wave Advantage QPU to avoid physical chain breaks.
Comparison of qubit requirements across different methods, highlighting the drastic reduction achieved by QPHR-ADMM.

Performance and Accuracy

The quantum-assisted solution achieved a total cost of $1,734.26k, representing only a 0.31% cost increase compared to the state-of-the-art classical solver, Gurobi. Furthermore, the PHR-ALM method proved more robust than basic penalty methods, which often struggle to converge to the true ground state on real quantum hardware due to environmental noise.

Conclusion and Future Work

While current "wall-clock" time is still impacted by classical-quantum communication latency, the QPHR-ADMM framework proves that large-scale power system optimization is possible on today's quantum hardware. Future improvements in QPU connectivity (moving from Chimera to Pegasus or Zephyr lattices) will allow for even larger ADMM blocks, further closing the gap between quantum research and industrial application in fields like Transmission Expansion Planning and Robust Unit Commitment.

Summary updated

Abstract

Stochastic Unit Commitment (SUC) has been proposed to manage the uncertainties driven by the integration of renewable energy sources. When solved by Benders Decomposition (BD), the master problem becomes a binary integer programming which is NP-hard and computationally demanding for classical computational methods. Quantum Annealing (QA), known for efficiently solving Quadratic Unconstrained Binary Optimization (QUBO) problems, presents a potential solution. However, existing quantum algorithms rely on slack variables to handle linear binary inequality constraints, leading to increased qubit consumption and reduced computational efficiency. To solve the problem, this paper introduces the Powell-Hestenes-Rockafellar Augmented Lagrangian Multiplier (PHR-ALM) method to eliminate the need for slack variables so that the qubit consumption becomes independent of the increasing number of bender's cuts. To further reduce the qubit overhead, quantum ADMM is applied to break large-scale SUC into smaller blocks and enables a sequential solution. Consequently, the Quantum-based PHR-ADMM (QPHR-ADMM) can significantly reduce qubit requirements and enhancing the applicability of QA in SUC problem. The simulation results demonstrate the feasibility of the proposed QPHR-ADMM algorithm, indicating its superior time efficiency over classical approaches for large scale QUBO problems under the D-Wave QPU showcases.
BibTeX
@misc{hong2025qubit,
  author = {Wei Hong and Wangkun Xu and Fei Teng},
  title = {{Qubit-Efficient Quantum Annealing for Stochastic Unit Commitment}},
  year = {2025},
  eprint = {2502.15917},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2502.15917},
}
Figure