Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Advancing Hybrid Quantum Neural Network for Alternative Current Optimal Power Flow

Ze Hu, Ziqing Zhu, Linghua Zhu, Xiang Wei, Siqi Bu, Ka Wing Chan

TL;DR

The paper addresses the computationally challenging AC-OPF problem by introducing a hybrid classical-quantum deep learning framework that leverages parameterized quantum circuits for feature extraction. It couples two residual learning schemes to mitigate barren plateaus and a physics-informed layer to enforce KKT constraints, enhancing physical consistency. Empirical results on IEEE 14-, 39-, and 162-bus systems show superior accuracy, generalization, and robustness to quantum noise, while using a small number of qubits. This work demonstrates the practical potential of NISQ-era quantum components for critical power-system optimization tasks and advances integration of quantum and physics-informed learning in real-time operations.

Abstract

Alternative Current Optimal Power Flow (AC-OPF) is essential for efficient power system planning and real-time operation but remains an NP-hard and non-convex optimization problem with significant computational challenges. This paper proposes a novel hybrid classical-quantum deep learning framework for AC-OPF problem, integrating parameterized quantum circuits (PQCs) for feature extraction with classical deep learning for data encoding and decoding. The proposed framework integrates two types of residual connection structures to mitigate the ``barren plateau" problem in quantum circuits, enhancing training stability and convergence. Furthermore, a physics-informed neural network (PINN) module is incorporated to guarantee tolerable constraint violation, improving the physical consistency and reliability of AC-OPF solutions. Experimental evaluations on multiple IEEE test systems demonstrate that the proposed approach achieves superior accuracy, generalization, and robustness to quantum noise while requiring minimal quantum resources.

Advancing Hybrid Quantum Neural Network for Alternative Current Optimal Power Flow

TL;DR

The paper addresses the computationally challenging AC-OPF problem by introducing a hybrid classical-quantum deep learning framework that leverages parameterized quantum circuits for feature extraction. It couples two residual learning schemes to mitigate barren plateaus and a physics-informed layer to enforce KKT constraints, enhancing physical consistency. Empirical results on IEEE 14-, 39-, and 162-bus systems show superior accuracy, generalization, and robustness to quantum noise, while using a small number of qubits. This work demonstrates the practical potential of NISQ-era quantum components for critical power-system optimization tasks and advances integration of quantum and physics-informed learning in real-time operations.

Abstract

Alternative Current Optimal Power Flow (AC-OPF) is essential for efficient power system planning and real-time operation but remains an NP-hard and non-convex optimization problem with significant computational challenges. This paper proposes a novel hybrid classical-quantum deep learning framework for AC-OPF problem, integrating parameterized quantum circuits (PQCs) for feature extraction with classical deep learning for data encoding and decoding. The proposed framework integrates two types of residual connection structures to mitigate the ``barren plateau" problem in quantum circuits, enhancing training stability and convergence. Furthermore, a physics-informed neural network (PINN) module is incorporated to guarantee tolerable constraint violation, improving the physical consistency and reliability of AC-OPF solutions. Experimental evaluations on multiple IEEE test systems demonstrate that the proposed approach achieves superior accuracy, generalization, and robustness to quantum noise while requiring minimal quantum resources.

Paper Structure

This paper contains 24 sections, 18 equations, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background
  3. Related works
  4. Research Gaps and Contributions
  5. Paper Organization
  6. Problem formulation
  7. Design of Hybrid Quantum Neural Network for AC-OPF
  8. Fundamentals of Quantum Neural Networks
  9. Hybrid Quantum Neural Networks
  10. Residual Learning (ResNet)
  11. Sequential Residual Connection
  12. Nested Residual Connection
  13. Physics-informed Neural Network
  14. Composite Noise Model
  15. Depolarizing Noise
  16. ...and 9 more sections

Figures (7)

  • Figure 1: Illustration of a qubit
  • Figure 2: Structure of a Hybrid QNN using residual connection and physics-informed layer
  • Figure 3: Structure of Hybrid QNN with two types of residual connections
  • Figure 4: Detailed illustration of a Hybrid Quantum PINN
  • Figure 5: Training loss curves for comparable methods on the IEEE 39-bus system
  • ...and 2 more figures