Counting Bases from Number of Qubits: Inferring VRP from Quantum Circuits

The security and confidentiality of sensitive information processed by quantum computers are of paramount importance, especially given quantum computers' potential to efficiently solve classically-hard optimization problems. At the heart of many transport optimization tasks lies the Vehicle Routing Problem (VRP), a complex combinatorial optimization problem classified as NP-hard. However, a promising avenue for approximating solutions to VRP is found in the Quantum Approximate Optimization Algorithm (QAOA). This paper demonstrates how leaking and learning simple parameters from QAOA quantum circuit structures, enables attackers to learn about the problem being solved. In routing optimization scenarios used by the military, for example, the attacker can learn location or connection of military bases. By exploiting information leakage during the QAOA execution, attackers can potentially breach security and retrieve sensitive VRP details, posing profound implications for civilian and national security.