Hybrid quantum–classical optimization · Quantum–GenAI co-design · Hardware-aware research
SMU-Quantum is an open research organization at the School of Computing and Information Systems, Singapore Management University (SMU).
Led by Prof. Hoong Chuin Lau, Professor of Computer Science at SMU and Senior Principal Scientist at A*STAR’s Institute of High Performance Computing, we study how quantum, quantum-inspired, classical, and learning-based methods can work together on difficult decision problems.
Our repositories contain research software, benchmark instances, tutorials, experiment artifacts, and hardware-oriented workflows. The work is grounded in practical constraints: limited qubits, circuit depth, noise, finite shots, penalty design, scalability, and strong classical baselines.
| Research direction | What we study |
|---|---|
| Quantum optimization | VQE, QAOA, CVaR-VQE, QRAO, PCE, QUBO/Ising mappings, and variational methods for constrained combinatorial problems. |
| Constraint and penalty design | Slack-free formulations, custom nonlinear penalties, finite-sampling objectives, CVaR-assisted optimization, feasibility-preserving encodings, and augmented Lagrangian methods. |
| Learning-augmented optimization | Reinforcement learning for penalty control, graph shrinking, decomposition, repair, multiplier updates, and adaptive solver policies. |
| Quantum reinforcement learning | Equivariant quantum circuits, size-invariant policies, cross-size transfer, TSP/QRL evaluation, and the limits imposed by simulation, finite shots, and hardware noise. |
| Quantum–GenAI co-design | LLM-guided closed-loop experimentation that searches over solver families, ansätze, optimizers, sampling budgets, compression strategies, and execution policies. |
| Noise mitigation and hardware execution | Light-cone cancellation, hardware-aware backend selection, circuit-resource analysis, fidelity diagnostics, and real-device benchmarking. |
| Use-inspired operations research | Vehicle routing, procurement, inventory, supply-chain resilience, maritime networks, finance, and resource planning. |
Near-term quantum computing is not only an algorithm-design problem. It is also a problem of representation, decomposition, learning, execution, and measurement.
We ask when a quantum component is useful, how to make it resource-aware, and how to evaluate it honestly against capable classical alternatives. We do not treat the use of a quantum circuit as evidence of quantum advantage.
| If you want to… | Start with… |
|---|---|
| Learn quantum optimization | Quantum Optimization Algorithms |
| Run benchmark experiments | Quantum Optimization Benchmarks |
| Explore adaptive quantum–classical control | AutoQResearch |
| Study vehicle-routing workflows | Adaptive Quantum CVRP |
| Investigate qubit-efficient encodings | Pauli Correlation Encoding |
| Study slack-free constraints | Cutting Slack |
| Browse the full publication record | Prof. Lau’s SMU profile |
-
From Circuits to Hardware: Benchmarking Standard and Qubit-Efficient Quantum Optimization on Real Hardware — accepted in Quantum Science and Technology. Benchmarks VQE, CVaR-VQE, QAOA variants, PCE, and QRAO across MDKP, MIS, QAP, and MSP on IBM Heron processors, covering 247 method–instance combinations. (DOI · Code and artifacts)
-
CVaR-Assisted Custom Penalty Function for Constrained Optimization — combines slack-free nonlinear penalties, finite sampling, and CVaR optimization for constrained binary problems, with experiments on Quantinuum H2 hardware.
-
Light Cone Cancellation for Variational Quantum Eigensolver in Solving Noisy Max-Cut — published in Scientific Reports. Develops a light-cone cancellation framework for reducing noise effects in VQE-based Max-Cut experiments.
-
AutoQResearch: LLM-Guided Closed-Loop Policy Search for Adaptive Variational Quantum Optimization — accepted as a QCE26 Quantum–GenAI Co-Design & Co-Discovery technical paper. Searches adaptive solver-control policies for MIS and CVRP instead of selecting one static solver configuration. (Repository)
-
Qubit-Scalable CVRP via Lagrangian Knapsack Decomposition and Noise-Aware Quantum Execution — decomposes CVRP into bounded-width per-vehicle knapsack subproblems, learns multiplier updates, and treats backend and circuit selection as part of the optimization loop. The paper explicitly does not claim quantum advantage.
-
Learning-Based Graph Shrinking for Quantum Optimization of Constrained Combinatorial Problems — in Quantum Computing and Artificial Intelligence: QC+AI 2026 (Springer CCIS 2872). Uses reinforcement learning and graph neural networks to guide graph shrinking before quantum solving. (SMU research record)
-
Understanding the Nature of Depth-1 Equivariant Quantum Circuit — introduces Size-Invariant Grid Search for analyzing and scaling depth-1 equivariant quantum reinforcement-learning circuits for TSP.
-
Diagnosing Simulation and Hardware Barriers to Cross-Size Transfer in Equivariant Quantum Reinforcement Learning — evaluates cross-size policy transfer across statevector simulation, tensor-network simulation, noisy execution, and hardware, while documenting the limits of scaling claims.
-
Securing the Flow: Maritime Energy Resilience under Correlated and Decision-Dependent Disruptions — develops a stochastic multi-commodity flow model with decision-dependent disruption probabilities, mean-CVaR risk, and Benders decomposition for maritime energy resilience.
-
Hybrid Learning and Optimization Methods for Solving the Capacitated Vehicle Routing Problem — combines Soft Actor-Critic reinforcement learning with augmented Lagrangian optimization for adaptive penalty control. (Code · Springer chapter)
-
Cutting Slack: Quantum Optimization with Slack-Free Methods for Combinatorial Benchmarks — studies dual ascent, bundle methods, cutting planes, and augmented Lagrangian formulations for TSP, MDKP, and MIS. (Code)
-
Adaptive Graph Shrinking for Quantum Optimization of Constrained Combinatorial Problems — introduces constraint-aware shrinking, verification and repair, and adaptive correlation updates for hardware-limited optimization.
-
A Comparative Study of Quantum Optimization Techniques for Solving Combinatorial Optimization Benchmark Problems — compares VQE, CVaR-VQE, QAOA variants, PCE, and QRAO across MDKP, MIS, QAP, and MSP. (Algorithms · Benchmarks)
The papers below include Prof. Lau and coauthors affiliated with SMU’s School of Computing and Information Systems:
- Solving Constrained Combinatorial Optimization Problems with Variational Quantum Imaginary Time Evolution — applies VarQITE to the Multiple Knapsack Problem and compares it with QAOA-based approaches.
- Implementing Slack-Free Custom Penalty Function for QUBO on Gate-Based Quantum Computers — evaluates custom penalty functions that avoid additional slack qubits for constrained binary optimization.
- A Feasibility-Preserved Quantum Approximate Solver for the Capacitated Vehicle Routing Problem — uses a feasibility-preserving encoding and Quantum Alternating Operator Ansatz for CVRP. (arXiv)
- Quantum Relaxation for Solving Multiple Knapsack Problems — combines QRAO-style quantum relaxation with classical linear relaxation for constrained procurement optimization. (QCE DOI)
- Quantum-Enhanced Simulation-Based Optimization for Newsvendor Problems — uses quantum amplitude estimation and learned demand distributions for stochastic inventory optimization. (QCE DOI)
- Quantum Monte Carlo Methods for Newsvendor Problem with Multiple Unreliable Suppliers — studies risk-aware inventory decisions with quantum Monte Carlo and quantum amplitude estimation.
The organization currently lists nine public repositories, including the organization profile configuration.
| Repository | Focus |
|---|---|
| autoqresearch | LLM-guided closed-loop policy search, staged confirmation, MIS/CVRP studies, experiment logs, checkpoints, plots, and hardware-run tooling. |
| quantum-optimization-benchmarks | Benchmark instances and evaluation artifacts for quantum optimization, including real-hardware benchmarking. |
| quantum-optimization-algorithms | Notebook-based implementations and examples for quantum optimization algorithms. |
| adaptive_quantum_cvrp | Modular CVRP framework combining ALM, SAC reinforcement learning, classical subproblem solving, and Qiskit VQE. |
| pauli-correlation-encoding | PCE optimization experiments, QUBO-to-Max-Cut transformations, and examples for knapsack, Max-Cut, and MIS. |
| cutting_slack | Notebooks and experiments for slack-free and Lagrangian-based constraint handling. |
| Repository | Focus |
|---|---|
| AMSI_Winter_School_Quantum_Tutorial | Tutorial notebooks introducing quantum optimization. |
| nature-of-depth1-eqc | Repository for the depth-1 equivariant quantum circuit study; the repository notes that full code will be released after manuscript acceptance. |
| .github | Organization-level profile and GitHub configuration. |
We aim to make research inspectable and extensible through source code, benchmark instances, experiment logs, notebooks, configuration files, and citation metadata where available.
- Research code may be experimental and is not necessarily production-ready.
- Results depend on problem instances, encodings, simulators, hardware, noise, mitigation, and classical baselines.
- A result that uses a quantum algorithm is not, by itself, evidence of quantum advantage.
- Each repository has its own license and reuse requirements. Check its
LICENSEfile before using or redistributing code. - Please cite the associated paper and repository when building on the work.
We welcome research collaborations with students, academic groups, quantum-computing teams, and organizations working on difficult decision problems.
- Research: propose a benchmark, optimization problem, algorithmic idea, or hardware study.
- Software: open an issue with a reproducible example before making a substantial change.
- Education: use the tutorial and algorithm repositories as starting points for courses and independent projects.
- Industry: discuss logistics, supply-chain, inventory, procurement, finance, maritime, and other resource-planning applications.
For formal research inquiries, contact Prof. Hoong Chuin Lau.
- SMU-Quantum on GitHub
- All public repositories
- Singapore Management University
- School of Computing and Information Systems
- Prof. Hoong Chuin Lau — SMU Faculty Directory
- Hoong Chuin Lau — NQCH Researcher Profile
Open research from Singapore Management University · Check each repository for its license and citation instructions.