Resource allocation for network profit maximization in NOMA-based F-RANs: a game-theoretic approach