Stealth assessment strategy in distributed systems using optimal deep learning with game based learning