Quantum-based hybrid framework for adaptive congestion management in smart grids with stochastic renewable penetration

Abstract

The simultaneous integration of renewable energy sources (RES) into the conventional system has introduced significant stochastic volatility in modern power systems which leads to frequent transmission line congestion and voltage instability. Hybrid optimization strategies, with a combination of classical Reinforcement Learning (RL) and Particle Swarm Optimization (PSO) are available for congestion management, but they frequently suffer from the curse of dimensionality and premature convergence in large-scale networks such as the IEEE 118-bus system [20, 3]. This paper proposes a Quantum-Hybrid framework by using Variational Quantum Reinforcement Learning combined with Quantum-behaved PSO (VQRL-QPSO) to address these limitations. The proposed architecture leverages Variational Quantum Circuits (VQC) to map high dimensional grid states into a Hilbert space using Angle Encoding, which allows the RL agent to learn optimal control policies with significantly fewer parameters than classical deep networks [4]. Furthermore, the QPSO component utilizes a wave function based search mechanism via a Delta potential well model which again enables "quantum tunneling" to escape local optima that typically entrap classical heuristics [5]. We have considered IEEE 39-bus and 118-bus systems for the simulation and the results indicate that the VQRL-QPSO framework achieves approx 25% reduction in congestion costs with 48% acceleration in convergence time compared to conventional RL-PSO benchmarks. Additionally, the model maintains superior grid resilience and voltage stability under 50% renewable uncertainty. Our proposed model establishes the potential of quantum-hybrid algorithms as scalable, real-time tools for ensuring stability in the next generation of high-penetration smart grids.

Received: 24 January 2026

Accepted: 19 April 2026

Published: 03 May 2026