Quantum solver for hard BInary Quadratic problems

Financiers

The QBIQ project is focused on developing an exact solver for binary quadratic optimization (QBO) problems, which belong to the class of NP-hard problems and play an important role in many application areas. The core computational framework will be based on the Branch-and-Bound algorithm and will combine state-of-the-art approaches in high-performance and quantum computing. The solver will consist of three modules: a high-performance solver building on the existing MADAM and BiqBin tools; an innovative hybrid quantum-classical solver that delegates the most demanding numerical computations to quantum hardware; and a pure quantum solver designed to test the limits of current quantum devices on unconstrained instances (QUBO). The solver will be accessible through a unified user interface with an integrated AI module that automatically selects the most suitable approach, while the source code will be made publicly available.

The research activities will focus on developing new methods for efficiently computing tight upper and lower bounds, applying semidefinite programming, and optimizing parallelization strategies. Special attention will be given to solving large systems of linear equations and generating feasible near-optimal solutions, where the use of noisy intermediate-scale quantum (NISQ) devices will also be explored. The project aims to demonstrate how the combination of classical supercomputers and quantum computers can lead to improved results in solving challenging optimization problems and to provide the research community with an open-source tool for further advances in quantum-classical computing.

Project Partners

Rudolfovo – Science and Technology Center Novo mesto.