内容简介

本书主要介绍在集值量测情况下的多智能体同步控制。首先，针对集值量测系统，介绍了两种辨识算法——无截断辨识算法和递推投影辨识算法，并分析了算法的性质；其次，基于两种辨识算法研究了不同拓扑结构不同系统动态下的多智能体同步控制，包括无向拓扑结构下的集值多智能体系统的同步控制、有向拓扑结构下的集值多智能体系统的同步控制、集值线性多智能体的同步控制，针对不同问题设计了不同的控制算法，并分别给出了同步控制速度。本书可供从事控制科学与工程、运筹学与控制理论、系统理论等专业人员阅读，也可供高等院校相关专业的师生参考。

前言概述

Overthelastthreedecades,numerousresearchhasbeendoneonthecooperativecontrolofmulti-agentsystemsbecauseofitswideapplications.Inamulti-agentsystem,allagentsarerequiredtocooperatewithotherstoachieveagroupobjectiveandeachofthemcanonlygetthelocalinformationfromitsneighbors.Consideringaccountuncertainties,disturbances,modelingerrors,etc,thelocalinformationisalwaysset-valuedinrealworldproblems.Consequently,theagentcanonlyobtaintheset-valued
informationoftheirneighbors’states.Theagentcanonlyobtaintheinformationwhetherthestatesoftheneighborsbelongtoasetorsomesets.Howtousetheselocalandset-valuedinformationtoachieveagroupobjectiveisarealisticandmeaningfulproblem.Thisbookintroduceshowagentsuselocalandset-valuedinformationtoachieveagroupobjective--consensus--byreachinganagreementontheirstates.Usually,
theseagentshavetopologicaldistributioninrealisticmodel.Theundirectedtopologyanddirectedtopologyareconsideredinthisbook.Themulti-agentsystemisdiscrete.Thefirst-ordermulti-agentsystemandthelinearmulti-agentsystemareintroducedanddiscussed.Differentconsensusalgorithmsaredesginedfordifferentmulti-agentsystems.
Themainideaoftheconsensusalgorithmsistodesignconsensuscontrolbytheestimatedstatesofneighbors.Fortheestimation,Iintroduceanofflineidentificationalgorithmandanonlineidentificationalgorithm.Basedontheofflineandonlineestimationmethods,Idesigntwokindsoftheconsensusalgorithms:two-time-scaleconsensusalgorithmandrecursiveconsensusalgorithm.
Finally,theconvergenceandconvergencerateoftheidentificationandcontrolalgorithmsareanalyzed.Myobjectiveinwritingthisresearchbookistosummarizemyworkinconsensuscontrolwithbinary-valuedmeasurements.Thebookiswrittenforresearchersandengineersworkingincontrolscienceandengineering,sensingnetwork,communication,signalprocessing,andmathematical
statistic,etc.Thebookismainlytheauthor’sunderstandingofconsensuswithbinary-valuedmeasurements.Interestedauthorscanrefertosomeclassicarticlesonquantizedconsensusinthereferencesofthebook.

作者介绍

王婷，北京科技大学人工智能学院副教授。王婷一直从事控制理论与应用方面的研究，对现代控制理论有全面的了解。从研究生期间开始研究集值量测下的系统辨识与趋同控制问题，已经取得了一些有意义的研究成果。目前已完成SCIE期刊论文13篇，其中2篇发表在控制类国际顶级期刊《IEEETransactionsonAutomaticcontrol》上，3篇发表在控制类国际知名期刊《IEEETransactionsonSystems,Man,Cybernetics:Systems》和《InternationalJournalofRobustandNonlinearControl》,2篇发表在国内知名期刊《ScienceChinaInformationSciences》上。

目录

Introduction

1.1.Background

1.2.Reviewonconsensusofmulti-agentsystems

1.3.OutlinetothisbookPreliminaries

2.1.Vectorsandnorms

2.2.Probabilitytheory

2.3.SomeinequalitiesIdentificationofset-valuedsystems

3.1.Problemformulation

3.2.Identificationalgorithmwithouttruncation

3.2.1.Algorithmdesign

3.2.2.Propertiesofthealgorithm

3.2.3.Simulationresults

3.3.Recursiveprojectionidentificationalgorithm

3.3.1.Algorithmdesign

3.3.2.Propertiesofthealgorithm

3.3.3.Simulationresults

3.4.NotesConsensusofmulti-agentsystemswithbinary-valuedmeasurementsandundirectedtopology

4.1.Problemformulation

4.2.Two-time-scaleconsensusalgorithm

4.2.1.Stochasticapproximationbasedalgorithm

4.2.2.Constantgainbasedalgorithm

4.3.Recursiveprojectionconsensusalgorithm

4.3.1.Algorithmdesign

4.3.2.Propertiesofthealgorithm

4.3.3.Simulationresults

4.4.NotesConsensusofmulti-agentsystemswithwithbinary-valuedmeasurementsanddirectedtopology

5.1.Problemformulation

5.2.Stochasticapproximationbasedalgorithm

5.2.1.Algorithmdesign

5.5.2.Propertiesofthealgorithm

5.2.3.Simulationresults

5.3.Constantgainbasedalgorithm

5.3.1.Algorithmdesign

5.3.2.Propertiesofthealgorithm

5.3.3.Simulationresults

5.4.NotesConsensusoflinearmulti-agentsystemswithbinary-valuedmeasurements

6.1.Problemformulation

6.2.AlgorithmDesign

6.3.Propertiesofthealgorithm

6.4.Simulationresults

6.5.Notesreferences