Quantum-Enhanced Double DQN for Real-Time Bus Scheduling: A Comparative Study

Efficient real-time scheduling of bus operations remains a critical challenge in urban transit systems, particularly under uncertain traffic conditions and high passenger demand variability. Our existing Double Deep Q-Network (Double DQN) implementation has demonstrated promising results in learning optimal holding and stop-skipping policies for bus operations. However, as urban networks grow in complexity and require increasingly stringent performance standards, classical deep reinforcement learning (DRL) architectures face fundamental limitations in scalability, convergence speed, and sample efficiency.
Quantum computing offers a transformative paradigm that leverages quantum mechanical phenomena—superposition, entanglement, and interference—to process information in fundamentally novel ways. Quantum neural networks (QNNs) and variational quantum circuits (VQCs) demonstrate potential for exponential representational capacity and accelerated convergence in specific learning tasks. This research proposes to develop and evaluate a Quantum-Enhanced Double DQN (QE-DDQN) that integrates quantum computing components with our existing classical framework.

Faculty Supervisor:

Saeid Saidi

Student:

Partner:

The University of Melbourne

Discipline:

Engineering

Sector:

Education

University:

University of Calgary

Program: