Adding enough free variables can turn non-convex problems into convex ones.

Could you explain why this is true?