Reasoning about commitments and penalties for coordination between autonomous agents