本文選題：一致性　切入點(diǎn)：合作競(jìng)爭(zhēng)多智能體網(wǎng)絡(luò)　出處：《東南大學(xué)》2017年碩士論文

【摘要】：近年來(lái),由于通信技術(shù)的迅猛發(fā)展,多智能體系統(tǒng)分布式控制的研究受到了越來(lái)越多的關(guān)注.分布式控制研究中很多有趣的話題已經(jīng)吸引了來(lái)自各行各業(yè)的注意,這些議題包括一致性、信息控制和同步等等.其中,一致性的研究已有好幾十年的歷史,現(xiàn)在依然是多智能體網(wǎng)絡(luò)的一個(gè)焦點(diǎn)問(wèn)題。本文研究了帶通信時(shí)滯的合作競(jìng)爭(zhēng)多智能體網(wǎng)絡(luò)的一致性問(wèn)題.主要工作如下:第一章對(duì)于多智能體網(wǎng)絡(luò)的一致性和分布式控制的研究發(fā)展進(jìn)行概述,并在此基礎(chǔ)上分析其研究現(xiàn)狀及前景,導(dǎo)出本文的主要工作.第二章探討了帶通信時(shí)滯的合作競(jìng)爭(zhēng)多智能體網(wǎng)絡(luò)的一致性問(wèn)題.對(duì)于一個(gè)給定的無(wú)向拓?fù)渚W(wǎng)絡(luò),針對(duì)線性耦合和非線性耦合兩種情況分別給出了多智能體狀態(tài)變化的動(dòng)態(tài)模型.應(yīng)用穩(wěn)定性理論和一些相關(guān)圖論的知識(shí),發(fā)現(xiàn)當(dāng)合作競(jìng)爭(zhēng)多智能體網(wǎng)絡(luò)連通時(shí)其所有智能體狀態(tài)在不考慮正負(fù)的情況下能夠?qū)崿F(xiàn)一致,也就是說(shuō)智能體的最終狀態(tài)之間只存在正負(fù)號(hào)的差異,而大小是相同的.這種二部一致?tīng)顟B(tài)的解對(duì)于線性耦合的情形此處給出了精確的求解表達(dá)式,對(duì)于非線性耦合的情形則給出了隱式表達(dá)式.最后,通過(guò)幾個(gè)數(shù)值仿真,驗(yàn)證了結(jié)論的有效性.第三章研究了有向合作競(jìng)爭(zhēng)多智能體網(wǎng)絡(luò)的一致性問(wèn)題,這里同樣考慮了多智能體間傳遞信息時(shí)的通信延遲影響.相比于第二章中考慮的情況,此處多智能體間的信息流動(dòng)是有方向的,也就是說(shuō)所考察的網(wǎng)絡(luò)對(duì)象從無(wú)向拓?fù)渚W(wǎng)絡(luò)拓展到了有向拓?fù)渚W(wǎng)絡(luò).利用李雅普洛夫定理和Barbalat引理,發(fā)現(xiàn)若有向合作競(jìng)爭(zhēng)多智能體網(wǎng)絡(luò)是強(qiáng)連通的,那么它能夠?qū)崿F(xiàn)二部一致.對(duì)于結(jié)構(gòu)平衡的網(wǎng)絡(luò),本文進(jìn)一步研究了所考慮系統(tǒng)的牽引二部一致性問(wèn)題.我們發(fā)現(xiàn)通過(guò)牽引控制有向符號(hào)網(wǎng)絡(luò)中的任意一個(gè)結(jié)點(diǎn),該網(wǎng)絡(luò)都能夠?qū)崿F(xiàn)牽引二部一致性.最后,給出兩個(gè)數(shù)值仿真的例子來(lái)闡述主要結(jié)論的有效性.第四章對(duì)主要內(nèi)容進(jìn)行了總結(jié),并對(duì)未來(lái)將要研究的方向做了簡(jiǎn)單介紹.
[Abstract]:In recent years, with the rapid development of communication technology, more and more attention has been paid to the research of distributed control in multi-agent system.Many interesting topics in distributed control research have attracted attention from a variety of industries, such as consistency, information control and synchronization.Among them, the research of consistency has been for decades, and it is still a focus of multi-agent network.In this paper, we study the consistency of cooperative and competitive multi-agent networks with communication delay.The main work is as follows: in the first chapter, the research and development of consistency and distributed control in multi-agent networks are summarized, and based on this, the present research situation and prospects are analyzed, and the main work of this paper is derived.In chapter 2, we discuss the consistency of cooperative and competitive multi-agent networks with communication delay.For a given undirected topological network, the dynamic models of multi-agent state change are given for linear coupling and nonlinear coupling respectively.By using stability theory and some related graph theory, it is found that all agent states can be consistent without considering the positive or negative when the cooperative competing multi-agent network is connected.In other words, there are only positive and negative sign differences between the final states of the agent, and the size is the same.For the linear coupling case, the exact solution expression is given for this bipartite uniform state solution, and the implicit expression is given for the nonlinear coupling case.Finally, the validity of the conclusion is verified by several numerical simulations.In chapter 3, the consistency problem of multi-agent networks with cooperative competition is studied. The effect of communication delay on the transmission of information between multi-agents is also considered.Compared with the situation considered in Chapter 2, the flow of information between multi-agents is oriented, that is to say, the network object is extended from undirected topology to directed topology.By using the Lyapunov theorem and Barbalat Lemma, it is found that if the multi-agent network is strongly connected to the cooperative competition, then it can achieve bipartial consistency.For the network with structural balance, the traction bipartite consistency of the system under consideration is further studied in this paper.We find that any node in the directed symbolic network can achieve traction bipartite consistency by traction control.Finally, two numerical simulation examples are given to illustrate the validity of the main conclusions.Chapter four summarizes the main contents and introduces the future research direction.
【學(xué)位授予單位】：東南大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP18;O157.5

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 李曉萌,楊煜普,許曉鳴;基于獨(dú)立學(xué)習(xí)的多智能體協(xié)作決策[J];控制與決策;2002年01期

2 郭魯;蘇文明;;企業(yè)內(nèi)組織的多智能體論述[J];科技廣場(chǎng);2008年02期

3 周，

本文編號(hào)：1709108