Is Mamba Compatible with Trajectory Optimization in Offline Reinforcement Learning?