多智慧型個體之訊息階流網絡

Multi-Agent Information Cascading Networks

蕭人豪(Jen-Hao Hsiao)

指導教授 ： 陳光禎

國立臺灣大學/電機資訊學院/電信工程學研究所/碩士(2018年)

https://doi.org/10.6342/NTU201804318

摘要

多智慧型個體系統操作在個體之間複雜的互動狀態以達成集體的任務，尤其，在多個體順序決策之中，個體進行社群學習來擷取其他個體的決策之中顯示的有用的訊息，個體的先行者對於決策有很大的影響，並且可能觸發訊息階流。訊息階流是在刻劃這樣一種情況，理性的個體因為接受到過多佔優勢的社群決策，而使得該個體可能忽略私自擁有的訊息做出決策。此論文發展了一套訊息階流的理論框架，可以將各種決策因素，像是先驗知識、不確定性、網絡結構與效用，整合並簡化為單純的門檻值決策。此外，更分析與探討了各種社群網絡結構，包含完全連接網絡、隨機網絡、無尺度網絡與小世界網絡，對於訊息階流、系統設計與效能的影響。此研究顯示了，作為一種操縱多智慧型個體系統行為的分散式方法學，訊息階流網絡理論可以應用在各式社群與工程系統之設計與分析。

關鍵字

多智慧型個體系統 ； 順序決策 ； 貝葉斯決策 ； 社群學習 ； 社群網絡 ； 隨機網絡 ； 訊息階流 ； 行為操縱

並列摘要

Multi-agent system operates under the complex interactions among agents to accomplish collective tasks. Particularly, in multi-agent sequential decision making, where agents perform social learning to obtain useful information from other's decisions, agent's predecessors have substantial influence on the decision making and can possibly trigger an information cascade. Information cascade characterizes the situation where rational agents are overloaded with the predominant social decisions and can make decisions ignoring their own private information. The thesis develops a theoretical framework of information cascades with capability to incorporate and simplify decision factors such as prior knowledge, uncertainty, network topology and utility into simple threshold decisions. In addition, the effect of social network topologies on information cascades, system design and performance is studied and analyzed, including the fully connected network, Erdos-Renyi network, scale-free network and small-world network. The study shows that the theory of information cascading networks can be applied to the design and analysis of various social and engineering systems as a distributed methodology of steering multi-agent system behavior.

並列關鍵字

multi-agent system ； sequential decision making ； bayes decision ； social learning ； social networks ； random networks ； information cascade ； behavior steering

參考文獻

[1] K.-C. Chen, M. Chiang, and H. V. Poor, “From technological networks to social networks,” IEEE Journal on Selected Areas in Communications, vol. 31, no. 9, pp. 548–572, 2013.

[2] M. O. Jackson, Social and economic networks. Princeton university press, 2010.

[3] M. Chiang, Networked Life: 20 Questions and Answers. Cambridge University Press, 2012.

[4] A.-L. Barabási et al., Network science. Cambridge university press, 2016.

[5] J. D. Papastavrou and M. Athans, “Distributed detection by a large team of sensors in tandem,” IEEE Transactions on Aerospace and Electronic Systems, vol. 28, no. 3, pp. 639–653, 1992.