More Like this

1. Deep Deterministic Policy Gradients with Transfer Learning Framework in StarCraft Micromanagement

   https://doi.org/10.1109/EIT.2019.8833742  

   Xie, Dong; Zhong, Xiangnan (May 2019, 2019 IEEE International Conference on Electro Information Technology (EIT))

   This paper proposes an intelligent multi-agent approach in a real-time strategy game, StarCraft, based on the deep deterministic policy gradients (DDPG) techniques. An actor and a critic network are established to estimate the optimal control actions and corresponding value functions, respectively. A special reward function is designed based on the agents' own condition and enemies' information to help agents make intelligent control in the game. Furthermore, in order to accelerate the learning process, the transfer learning techniques are integrated into the training process. Specifically, the agents are trained initially in a simple task to learn the basic concept for the combat, such as detouring moving, avoiding and joining attacking. Then, we transfer this experience to the target task with a complex and difficult scenario. From the experiment, it is shown that our proposed algorithm with transfer learning can achieve better performance. 

   more » « less
2. Local Advantage Actor-Critic for Robust Multi-Agent Deep Reinforcement Learning.

   Yuchen Xiao, Xueguang Lyu (January 2021, the International Symposium on Multi-Robot and Multi-Agent Systems)

   Policy gradient methods have become popular in multi-agent reinforcement learning, but they suffer from high variance due to the presence of environmental stochasticity and exploring agents (i.e., non-stationarity), which is potentially worsened by the difficulty in credit assignment. As a result, there is a need for a method that is not only capable of efficiently solving the above two problems but also robust enough to solve a variety of tasks. To this end, we propose a new multi-agent policy gradient method, called Robust Local Advantage (ROLA) Actor-Critic. ROLA allows each agent to learn an individual action-value function as a local critic as well as ameliorating environment non-stationarity via a novel centralized training approach based on a centralized critic. By using this local critic, each agent calculates a baseline to reduce variance on its policy gradient estimation, which results in an expected advantage action-value over other agents’ choices that implicitly improves credit assignment. We evaluate ROLA across diverse benchmarks and show its robustness and effectiveness over a number of state-of-the-art multi-agent policy gradient algorithms. 

   more » « less
3. Contrasting Centralized and Decentralized Critics in Multi-Agent Reinforcement Learning

   Xueguang Lyu, Yuchen Xiao (January 2021, Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems)

   Centralized Training for Decentralized Execution, where agents are trained offline using centralized information but execute in a decentralized manner online, has gained popularity in the multi-agent reinforcement learning community. In particular, actor-critic methods with a centralized critic and decentralized actors are a common instance of this idea. However, the implications of using a centralized critic in this context are not fully discussed and understood even though it is the standard choice of many algorithms. We therefore formally analyze centralized and decentralized critic approaches, providing a deeper understanding of the implications of critic choice. Because our theory makes unrealistic assumptions, we also empirically compare the centralized and decentralized critic methods over a wide set of environments to validate our theories and to provide practical advice. We show that there exist misconceptions regarding centralized critics in the current literature and show that the centralized critic design is not strictly beneficial, but rather both centralized and decentralized critics have different pros and cons that should be taken into account by algorithm designers 

   more » « less
4. Value-Decomposition Multi-Agent Actor-Critics

   Su, Jianyu; Adams, Stephen; Beling, Peter (February 2021, Proceedings of the AAAI Conference on Artificial Intelligence)

   The exploitation of extra state information has been an active research area in multi-agent reinforcement learning (MARL). QMIX represents the joint action-value using a non-negative function approximator and achieves the best performance on the StarCraft II micromanagement testbed, a common MARL benchmark. However, our experiments demonstrate that, in some cases, QMIX performs sub-optimally with the A2C framework, a training paradigm that promotes algorithm training efficiency. To obtain a reasonable trade-off between training efficiency and algorithm performance, we extend value-decomposition to actor-critic methods that are compatible with A2C and propose a novel actor-critic framework, value-decomposition actor-critic (VDAC). We evaluate VDAC on the StarCraft II micromanagement task and demonstrate that the proposed framework improves median performance over other actor-critic methods. Furthermore, we use a set of ablation experiments to identify the key factors that contribute to the performance of VDAC. 

   more » « less
5. Value-decomposition multi-agent actor-critics

   Su, Jianyu; Adams, Stephen; Beling, Peter (February 2021, Proceedings of the AAAI Conference on Artificial Intelligence)

   The exploitation of extra state information has been an active research area in multi-agent reinforcement learning (MARL). QMIX represents the joint action-value using a non-negative function approximator and achieves the best performance on the StarCraft II micromanagement testbed, a common MARL benchmark. However, our experiments demonstrate that, in some cases, QMIX performs sub-optimally with the A2C framework, a training paradigm that promotes algorithm training efficiency. To obtain a reasonable trade-off between training efficiency and algorithm performance, we extend value-decomposition to actor-critic methods that are compatible with A2C and propose a novel actor-critic framework, value-decomposition actor-critic (VDAC). We evaluate VDAC on the StarCraft II micromanagement task and demonstrate that the proposed framework improves median performance over other actor-critic methods. Furthermore, we use a set of ablation experiments to identify the key factors that contribute to the performance of VDAC. 

   more » « less