Multi-Agent Reinforcement Learning and Genetic Policy Sharing
نویسنده
چکیده
The effects of policy sharing between agents in a multi-agent dynamical system has not been studied extensively. I simulate a system of agents optimizing the same task using reinforcement learning, to study the effects of different population densities and policy sharing. I demonstrate that sharing policies decreases the time to reach asymptotic behavior, and results in improved asymptotic behavior.
منابع مشابه
Intelligent Multi-Agent Dynamical Systems and Policy Sharing
The effects of policy sharing between agents in a multi-agent dynamical system has not been studied extensively. I simulate a system of agents optimizing the same task using reinforcement learning, to study the effects of different population densities and policy sharing. Rich behavior emerges, dependent on the varied parameters, giving insight towards the dynamics of the system. We demonstrate...
متن کاملParameter Sharing Deep Deterministic Policy Gradient for Cooperative Multi-agent Reinforcement Learning
Deep reinforcement learning for multi-agent cooperation and competition has been a hot topic recently. This paper focuses on cooperative multi-agent problem based on actor-critic methods under local observations settings. Multi agent deep deterministic policy gradient obtained state of art results for some multi-agent games, whereas, it cannot scale well with growing amount of agents. In order ...
متن کاملParallel Fault Tolerant Multi-Agent Reinforcement Learning
Reinforcement learning is a powerful tool for training an agent in a sequential decision based environment and has been successful in many simulated [6] as well as practical [5] domains. In this paper we investigate methods of strengthening the rate of convergence of a single agent RL learner by sharing observations with other independent agents. In contrast to multi-agent reinforcement methods...
متن کاملMulti-Agent Shared Hierarchy Reinforcement Learning
Hierarchical reinforcement learning facilitates faster learning by structuring the policy space, encouraging reuse of subtasks in different contexts, and enabling more effective state abstraction. In this paper, we explore another source of power of hierarchies, namely facilitating sharing of subtask value functions across multiple agents. We show that, when combined with suitable coordination ...
متن کاملReinforcement Learning Algorithms for Homogenous Multi-Agent Systems
We examine the situation in which multiple reinforcement learning agents attempt to speed up the learning process by sharing information. The agents do not interact with each other in the enviroment; instead each agent is learning the same task in isolation from the other agents. As reinforcement learning is a trial-and-error approach, an agent’s performance in the environment typically increas...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- CoRR
دوره abs/0812.1599 شماره
صفحات -
تاریخ انتشار 2008