Loan Portfolio Optimization with Variational Quantum Algorithms
π arXiv:2609.30195 Β· π₯ PDF Β· 2026-09-24 Β· quant-ph
Authors: Balaganchi A. Bhargava [arXiv Β· scholar] , Kamlesh Kumar [arXiv Β· scholar] , Gouranga Dinda [arXiv Β· scholar] , Aniruddha Biswas [arXiv Β· scholar] , Vijay S. Rao [arXiv Β· scholar] , Debarag Banerjee [arXiv Β· scholar] , Arun Sehrawat [arXiv Β· scholar]
π Abstract
Loan portfolio optimization (LPO) seeks portfolios that minimize credit risk while satisfying practical selection constraints. Accounting for portfolio-level risk requires modeling not only expected losses but also loss variability and borrower-default correlations, leading to a large-scale combinatorial optimization problem whose complexity grows rapidly with portfolio size. In this work, we formulate LPO as a Quadratic Constrained Binary Optimization (QCBO) problem that jointly captures expected and unexpected losses and transform the resulting formulation into a Quadratic Unconstrained Binary Optimization (QUBO) model suitable for variational quantum optimization. To address the limited qubit resources of current quantum hardware, we employ Pauli Correlation Encoding (PCE), which enables a large number of portfolio-selection variables to be represented using comparatively few qubits. The proposed framework is evaluated on a real-world microfinance dataset comprising 2012 borrowers. For problem instances with up to 1000 portfolio-selection variables, we benchmark classical simulations of a variational quantum algorithm (VQA) against Google OR-Tools under matched time budgets, obtaining feasible portfolios with an objective-value gap of about 40% across all sizes. We further implement the framework on superconducting quantum hardware for instances containing up to 1500 variables. Across these instances, the experimentally obtained objective values are within 10% of the corresponding simulation results. These findings demonstrate the feasibility of applying PCE-based variational quantum optimization to large-scale loan portfolio optimization under current quantum hardware limitations.