Scribd
Upload
15 views7 pages

Report On Consensus Problem in Networks of Agents

This report analyzes consensus problems for networks of agents with switching topologies and time-delays. It studies directed networks with fixed and switching topologies, as well as undirected networks with time-delays and fixed topologies. The analysis relies on tools from algebraic graph theory, matrix theory, and control theory to study how consensus is reached in different network configurations.

Uploaded by

Yufei Han
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
15 views7 pages

Report On Consensus Problem in Networks of Agents

This report analyzes consensus problems for networks of agents with switching topologies and time-delays. It studies directed networks with fixed and switching topologies, as well as undirected networks with time-delays and fixed topologies. The analysis relies on tools from algebraic graph theory, matrix theory, and control theory to study how consensus is reached in different network configurations.

Uploaded by

Yufei Han
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 7

Report on Consensus Problem in Networks of Agents With

Switching Topology and Time-delays

Many coordination control tasks can be united to the same theory structure-the consensus problem
of multi-agent networks. In many applications involving multi-agents systems, groups of agents
need to agree upon certain quantities of interest. Such quantities might or might not be related to
the motion of the individual agents. In the complex environment, the communication between
agents is often delayed, and the connection is not fixed due to link failures or creations in
networks with mobile nodes. So it is of great importance to study on consensus problems in
networks of agents with switching topology and time-delays.
In this paper, 3 cases of consensus problems for networks of dynamic agents with fixed and
switching topologies are analyzed:
1)Directed networks with fixed topology;
2)Directed networks with switching topology;
3)Undirected networks with communication time-delays and fixed topology;
This report focuses on the conclusion of the analysis. And at the end of the report, unresolved
problems which worth further study are listed out.

I. INTRODUCTON
Distributed coordination of networks of dynamic agents has been widely used and researched as
multi-agents gain their popularity. Groups of agents need to agree upon certain quantities of
interest, so it is important to address agreement problems in their general form for networks of
dynamic agents with directed information flow under link failure and creation. The convergence
analysis is provided for each case. And here, a common Lyapunov function is introduced to
guarantee asymptotic convergence to a group decision value in network with switching topology.
A connection between algebraic connectivity of the network and the performance of reaching an
agreement is established. The maximum time-delay that can be tolerated by a network of
integrators applying a linear consensus protocol is inversely proportional to the largest eigenvalue
of the network topology or the maximum degree of the nodes of the network. For undirected
network with fixed topology ,there is a tradeoff between performance of reaching a consensus and
robustness to time-delays and a second tradeoff between achieving a high performance and
maintaining a low communication cost.
The analysis relies on tools from algebraic graph theory, matrix theory, and control theory. A
connection between the performance of a linear consensus protocol on a directed network and the
Fiedler eigenvalue of the mirror graph of the information is built. A class of directed graph named
balanced graph play an important role in derivation of an invariant quantity , a Lyapunov function
for convergence analysis of average-consensus problem, and the analysis of the convergence of an
agreement protocol for networks with switching topology.

II. TO DEFINE CONSENSUS PROBLEMS ON GRAPHS


After the mathematical modeling, we get a set of nodes that represents physical quantities of the
agents, a weighted adjacency matrix whose elements stand for the links between the nodes, and
functions which describe their dynamical state. That is, we get a dynamic graph. The initial state
of the system is not fixed but a function of the value(x) of the nodes. A state feedback

is a distributed protocol, which can consensus problem of the system if

and only if there exists an asymptotically stable equilibrium that satisfies the network dynamics.
Plus, in a dynamic network with switching topology, the information flow is a discrete-state of the
system that changes in time.

III. CONSENSUS PROTOCOLS

a is the weight, x is the value of the node. Choosing a is an optimization problem that
falls within the category of network design problems. In each case, consensus is asymptotically
reached.
The communication cost of protocols with directed information flow is smaller than the
communication cost of their undirected counterparts, and the information floe in a flock id
directed and the topology of the network of agents goes through changes that are discrete-event
type in nature, those are reasons why we analyze consensus protocols for digraphs.

Notation : except part X, the rest parts are all based on protocol(A1).

IV. NETWORK DYNAMICS AND THEIR RELATION TO GRAPH LAPLACIANS IF


PROTOCOL(A1) IS GIVEN
For a network of continuous-time integrator agents,
The stability properties of system(8) depends on the eigenvalue of the graph Laplacian L.
In a network with switching topology, the convergence analysis is as followed,

For agents with discrete models, the network dynamics is

V. SOME BACKGROUND ON ALGEBRAIC GRAPH THEORY AND MATRIX THEORY


RELATED TO THE PROPERTIES OF GRAPH LAPLACIANS

VI. THERE EXISTS A STRONGLY CONNECTED DIGRAPH THAT DOES NOT SOLVE AN
AVERAGE-CONSENSUS PROBLEM.

VII. DIRECTED NETWORKS WITH FIXED TOPOLOGY AND BALANCED GRAPHS

Theorem 4: Consider a network of integrators with a fixed topology that is a strongly


connected digraph. Then, protocol (A1) globally asymptotically solves the average-
consensus problem if and only if G is balanced.
Theorem 5: Consider a network of integrator agents with a fixed topology that is a
strongly connected digraph. Then, protocol (A1) globally asymptotically solves the
average-consensus problem if and only if .
Corollary 2(works even if the graph is unbalanced): Assume all the conditions in Theorem5 hold.
Suppose L has a nonnegative left eigenvector associated with that
satisfies . Then, after reaching a consensus, the group decision value is

Corollary 3. (Multirate Integrators): Consider a network of multirate integrators with the node
dynamics

Assume the network has a fixed topology and each node applies protocol (A1).

Then, an agreement is globally asymptotically reached and the group decision value will be

In theorem6, graph G does not need to be connected. And we can get the conclusion that for
networks with balanced information flow, is an invariant quantity. This will be used in
the following derivation.

VIII. MIRROR GRAPHS AND THEIR USE ON DETERMINING THE SPEED OF


CONVERGENCE OF A CONSENSUS PROTOCOL ON DIGRAPHS AND DEFINING THE
ALGEBRAIC CONNECTION OF DIGRAPHS
A well-known observation regarding the Fiedler eigenvalue of an undirected graph is that for
dense graphs is relatively large and for sparse graphs is relatively small. So a network with
dense interconnections solves an agreement problem faster than a connected but sparse network.

IX.NETWORKS WITH SWITCHING TOPOLOGY


In terms of the network G, switching topology means certain number of edges are added or
removed from the graph. However, it is still possible to reach a consensus.
X. NETWORKS WITH COMMUNICATION TIME-DELAYS
We focus on the simplist possible case where the time-delays in all channels are equal to
and the network topology is fixed and undirected.

A.Tradeoff Between Performance and Robustness


,

B. Tradeoff Between High Performance and Low Communication Cost


For undirected graphs with 0–1 weights, a graph with a relatively high communication cost is
expected to have a relatively high algebraic connectivity .

XI. EXTENDED PROBLEMS


With the development of control theories in the past decade, this masterpiece of its day shows
some limitations .

1)When put into use, multi-agent systems can hardly avoid the influence of the environment, so
the time-delay,the noise and external disturbance should all be taken into consideration. While in
this paper, the noise is not the main concern.
2)This paper studies on linear consensus protocol and think little about nonlinear system.
3) It would be more practical if the time it takes is worked out ，while this paper just focuses on
whether we can reach to the consensus or not with switching topology and time-delays.

XII. STATEMENT ON THE REPORT


This report was finished on July 27th. I should have delivered it earlier. But it is not delivered until
now because I was at home then and had no authoritative books to refer to. So I bought the most
recent train ticket and went back to school. However, the most useful thing I found in the library
was details on Lyapunov function. I am really upset about postponing delivering my report
because after all it is not better polished.
I didn’t list the proof of theorems required in the paper, for after trying I find that it will make this
report too obscure and repetitive and long-winded. So finally I choose this way to show what I get
from the report clearly.
This is my first report on an IEEE paper, please forgive me for my hesitation and lack of
confidence. And thank you for reading it.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy