arXiv.wiki

Training variational quantum algorithms is NP-hard -- even for logarithmically many qubits and free fermionic systems

Lennart Bittel, Martin Kliesch

https://arxiv.org/abs/2101.07267 (or view on SciRate)


Main idea

Implications

Four settings showing hardness

  1. oracle: quantum computer; proving the optimization is NP-hard even if constant relative precision
  2. ground state problem is easy; log(n) qubits
  3. ground state problem is easy; free fermions
  4. ground state problem is easy; QAOA setup

Details



tags: QAOA  quantum-complexity  variational-quantum-algorithms