It's difficult to trust self-published, non-peer reviewed papers on arxiv. You can just as easily find papers claiming P=/=NP
That really sounds suspicious to me. Why not just solve this version directly with these changes. That should have been a more straightforward proof.