高粒度自适应代理软件社会？
我在想我们的社会如何激励一群自由个体的行为，使这些个体的工作满足整个社会的要求。 例如，人们通过工作做出贡献，赚钱，然后用这些钱来获得他人贡献的公平份额。 它在大多数情况下都有效……但是你需要警察来防止人们偷窃等，这使得人们在不做出贡献的情况下不公平地分享他人的贡献。
马文·明斯基（Marvin Minsky）有一本书《思想社会》，这本书可能沿着这些方向发展了一些想法。 他们是神经网络中关于紧急行为、细胞自动机等的较旧的想法。最近有关于合作代理等的工作。
但是，我认为没有人设计过一个大型的高粒度代理系统，旨在作为一个整体系统工作和发展（即稳态的不同角色，如警察等）。 您知道有类似的工作吗？
I was thinking about how our society incentivizes behaviors by a collection of free individuals, so that the work of those individuals fulfils the requirements of the society as a whole. For example, people contribute by working, earn money, and use that money to gain a fair share of the contributions of others. It works for the most part ... but you need police to prevent people from stealing, etc. which allows people to unfairly share in the contributions of others without contributing.
Marvin Minsky had a book "Society of Minds" which may have developed some ideas along these directions. And their were the older ideas in neural nets about emergent behavior, cellular automata, etc. And recently there is work on cooperative agents etc.
However, I don't think anyone has designed a large system of high-granularity agents that are designed to work and evolve as a total system (i.e. different roles for homeostatis like police etc.). Do you know of any work along these lines?
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）
Miller 和 Drexler 的agoric 论文介绍了这一研究领域。 “激励工程：计算资源管理”最具技术含量； 已经有一些计划实现了这一愿景的某些方面，但据我所知，还没有整个计划。
Miller and Drexler's agoric papers introduced this field of study. "Incentive Engineering: for Computational Resource Management" has the most technical meat to it; there have been some programs realizing aspects of this vision but not yet the whole package AFAIK.
点对点网络不就是这样的吗？
除此之外，现在出现了一些非常关注代理的计算机语言：Clojure、Scala、Erlang、Go……
Aren't peer-to peer networks something like that?
Apart of that, there are now apearing computer languages with strong focus on agents: Clojure, Scala, Erlang, Go...