Comparative Evaluation of Multi-Agent Deep Reinforcement Learning Algorithms