arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2207.07388cs.MAcs.AI
收藏

随机市场博弈

Stochastic Market Games

LMU Munich(慕尼黑大学);Technische Hochschule Ingolstadt(英戈尔施塔特应用技术大学)

Kyrill Schmid, Lenz Belzner, Robert Müller, Johannes Tochtermann, Claudia Linnhoff-Popien

摘要:

多智能体系统的一些最相关的未来应用,如自动驾驶或工厂即服务,显示混合动机的场景,其中智能体可能有相互冲突的目标。在这些环境中,代理很可能在独立学习下的合作方面学习到不期望的结果,例如过度贪婪的行为。从现实世界的社会,在这项工作中,我们建议利用市场力量提供激励代理成为合作。正如囚徒困境的迭代版本中所展示的那样,拟议的市场制定可以改变游戏的动态,以不断学习合作政策。此外,我们评估我们的方法在空间和时间上扩展设置不同数量的代理。我们的实证研究发现,市场的存在可以提高整体的结果和代理人的个人收益,通过他们的交易活动。

英文摘要:

Some of the most relevant future applications of multi-agent systems like autonomous driving or factories as a service display mixed-motive scenarios, where agents might have conflicting goals. In these settings agents are likely to learn undesirable outcomes in terms of cooperation under independent learning, such as overly greedy behavior. Motivated from real world societies, in this work we propose to utilize market forces to provide incentives for agents to become cooperative. As demonstrated in an iterated version of the Prisoner's Dilemma, the proposed market formulation can change the dynamics of the game to consistently learn cooperative policies. Further we evaluate our approach in spatially and temporally extended settings for varying numbers of agents. We empirically find that the presence of markets can improve both the overall result and agent individual returns via their trading activities.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑