QUBO-Based Optimization of Hybrid Renewable Energy Portfolios: Benchmarking Classical and Quantum-Inspired Solvers

Author: Devavrat J. Gokal

Affiliation: Hill Spring International School, Mumbai

Abstract:

This paper considers the problem of optimal site selection for hybrid solar-wind energy systems by formulating it as a Quadratic Unconstrained Binary Optimization (QUBO) model, which seeks to maximize annual energy yield under budgetary and site-exclusivity constraints. The performance of a classical metaheuristic solver (MQLib) is benchmarked against a quantum-inspired solver (D-Wave’s Simulated Annealing) using empirical solar and wind resource data across up to 60 candidate sites in Gujarat and Maharashtra, India. Results show that D-Wave’s simulated annealing consistently yields higher objective values and exhibits more favorable computational scaling as the problem size increases. In addition, a related location-allocation problem for energy system planning, previously optimized by Ajagekar and You in 2019 using Gurobi and the D-Wave 2000Q is revisited. The results demonstrate that MQLib outperforms both solvers across all benchmark instances in terms of solution quality and computational runtimes. The relative strengths and limitations of classical and quantum-inspired solvers are examined, with emphasis on trade-offs in scalability, solution quality, and real-world applicability. While D-Wave’s current simulated annealing shows better performance in this setting, this advantage may diminish with the adoption of newer hardware such as the D-Wave Advantage2 system, which is expected to offer higher qubit connectivity and improved noise performance.

Keywords: QUBO, Combinatorial Optimization, Renewable Energy, Metaheuristic Algorithms, Quantum Computing

ISSN 3069-8200

Download Full Article (PDF)