https://arxiv.org/pdf/quant-ph/0007071.pdf
In Chapter 6 in the paper "A Numerical Study of the Performance of a Quantum Adiabatic Evolution Algorithm for Satisfiability”, by Edward Farhi, Jeffrey Goldstone and Sam Gutmann, it says that if we destroy the bit structure, then AQC on the exact cover problem does not work as well.
I don't quite see why, especially intuitionally. If we only permute the qubits (if I understood the article correctly), why would AQC not work?
Could someone help me understand please?
