多代理系统只是炒作吗？
作为一名研究人员，我很想听听人们对多代理系统的看法，如果您当然想到了整个想法的话。您是否认为这不仅仅是炒作和另一个流行语？您能看到商业或日常计算中的任何潜在用途吗？或者您是否认为我们已经可以通过简单、更优雅的解决方案实现 MAS 所提供的一切？
As a researcher I am curious to hear what people think of Multi-Agent Systems if of course you came across the whole idea. Do you believe there is something more in there than just a hype and another buzzword? Can you see any potential uses in business or everyday computing? Or do you think that we can already achieve everything MAS has to offer but with simple more elegant solutions?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
我是一名研究教授，在自治代理和多代理系统会议（AAMAS）上发表了许多文章：多代理研究的主要工具。
MAS 是研究人员使用的术语（于 1995 年左右为第一届多代理系统国际会议创造） （ICMAS），它将分布式人工智能（DAI）和自治代理研究社区聚集在一个帐篷下：MAS帐篷），指的是组织自治代理团队的算法和方法。 MAS 的研究人员开发了机器人足球算法（参见 Robocup），协调自主机器人漫游车（如火星漫游车） 、分布式资源分配（谁执行哪个任务）以及许多其他领域。
我不认为有任何你所描述的“炒作”。您可以阅读所有过去会议中的论文，每一篇都清楚地说明了作者试图完成的任务，他如何尝试，以及结果是什么。我不知道有人对这些技术的力量做出愚蠢的主张：它们都只是算法（并不是一切）。这里没有魔法。
问题
不正确的是，如果您可以使用简单且更优雅的解决方案来解决 MAS 问题，那么您的解决方案现在就是 MAS 解决方案！
MAS 是一个问题域，以及我们迄今为止找到的解决方案。如果您找到更好的解决方案，那就太棒了，发布它并加入 MAS 社区。
顺便说一句，我经常看到这种混乱。熟练程序员没有意识到研究社区（通常）是由他们所解决的问题而不是解决方案来定义的。
I am a research professor who has published many articles in the Autonomous Agents and Multiagent Systems Conference (AAMAS): the main vehicle for multiagent research.
MAS is a term used by researchers (coined around 1995, for the first International Conference on Multiagent Systems (ICMAS), which brought together the Distributed Artificial Intelligence (DAI) and the Autonomous Agents research communities under one tent: the MAS tent) that refers to algorithms and methods for organizing teams of autonomous agents. Researchers in MAS have developed algorithms for Robot soccer (see Robocup), coordinating autonomous robot rovers (as in Mars rovers), distributed allocation of resources (who does which task), and many other domains.
I don't see that there is any "hype" as you describe. You can read all the papers in past conferences and each one clearly states what the author tried to accomplish, how he tried it, and what the results were. I do not know of anyone making silly claims about the power of these techniques: they are all just algorithms (isn't everything). No magic here.
The question
is incorrect in that, if you can solve a MAS problem with a simple and more elegant solution, your solution is now a MAS solution!
MAS is a problem domain, along with the solutions we found so far. If you find a better solution then, awesome, publish it and join the MAS community.
As an aside, I see this confusion often. Journeymen programmers don't realize that research communities are (usually) defined by the problem they work on, not a solution approach.
与人工智能和技术的许多其他领域相比，多智能体系统的宣传还不够！
我发现有些人对多智能体系统一无所知，却活跃在技术、编程和“人工智能”领域。 （引用它现在是炒作，实际上已经失去了所有意义）
我在 2008 年通过 Netlogo 了解了多智能体系统，它改变了我对使用计算技术解决问题的看法。我当时意识到这些类型的程序需要不断增强的计算能力。最近，我了解了所有炒作驱动的东西（数据科学、ML、DNN、RL 等），我认为所有这些炒作都将与 MAS 很好地集成，但尚未得到充分理解。许多人是通过 MMO 游戏了解到这一想法的，该游戏大受欢迎，因此可能还会有飞跃。
Compared to many other fields of Artificial Intelligence and Technology, multi-agent systems aren't hyped enough!
I find people who know nothing about Multi-agent systems and are actively in the field of technology, programming, and "artificial intelligence". (quoted since it now is hype, and has effectively lost all meaning)
I learned about Multi-agent systems in 2008 through Netlogo, and it changed my perspective about problem solving using computational technology. I realized at the time that these types of programs would require increasing computing power. More recently I have learned all the hype driven stuff (data-science, ML, DNN, RL, etc.) I think all this hype will integrate well with MAS, and has yet to be fully understood. Many people are introduced to this thought through MMO gaming, which has been a huge hit, so there may be a leap yet to come.