Reverse Engineering of Hamiltonian Expressions from D-Wave programs

Quantum annealing software has achieved a certain market penetration since it has demonstrated a good performance for optimization problems. The problem definitions (Hamiltonians functions to be minimized) are usually built in combination with classical software that can evolve over time. Thus, it is difficult to comprehend the underlying Hamiltonian and represent it in an accurate way for maintaining such programs. Although some reverse engineering techniques have been proposed for gate-based software, there are no choices to accomplish reverse engineering in quantum annealing. This research proposes a dynamic analysis technique of D-Wave (python) programs for reversing Hamiltonians, which has been preliminarily evaluated with nine programs. The technique also represents those expressions by using the Knowledge Discovery Metamodel (ISO/IEC 19506). Due to the usage of this standard, the outgoing expressions can be represented in combination with other parts of classical-quantum software systems. Thus, the main implication is that this technique contributes to the software modernization of/towards hybrid information systems.