0% found this document useful (0 votes)
8 views25 pages

Stream Project

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
8 views25 pages

Stream Project

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 25

Department of Mathematical Sciences

Stream Project (Sem-V)


Fixed Time Consensus Analysis of Multi
Agent Systems via Impulsive Control

Presented by Debangi Ghosh


Supervised by Dr. Santwana Mukhopadhyay

Date- 20.11.2024
Acknowledgement
I would like to express my heartfelt
gratitude to Dr. Santwana
Mukhopadhyay Ma’am for her invaluable
guidance, timely feedback, and
unwavering support throughout the
course of this project. Her insights,
coupled with the research papers and
relevant materials she provided, have
been instrumental in shaping the
direction of this work. It is my privilege to
have her guidance as I continue this
journey toward completing the project.
INTRODUCTION
Multi-agent systems are systems that
consist of multiple interacting intelligent
agents that operate in a shared
environment to achieve collective goals.
They are advantageous over single agent
systems for the following reasons- (1)
lowering task costs; (2) implementing
parallel processing; (3) achieving stability
and reducing error; (4) being scalable;
and, (6) distributed decision making.
Achieving consensus in Multi-agent
systems is a topic of great interest.
NON LINEAR DYNAMICAL SYSTEMS
A mathematical model of a
system’s evolution across time is
called a dynamical system. It is Multiple equilibrium points
composed of a number of variables
that indicate the system’s current
state and a collection of rules, or Limit cycles and oscillations
equations, that control how these
variables change between states.
Non-linearity in dynamical systems Bifurcations
can arise from various sources,
such as non-linear terms in the
governing equations, non-linear Chaos and sensitivity to conditions
interactions between system
components, or external forcing
functions.
NON LINEAR DYNAMICAL SYSTEMS
The following set of coupled first-order differential equations can be used to
model a general non-linear system.

Non-linear dynamic systems are of two kinds:


autonomous and non-autonomous.
represents autonomous systems.
represents non-autonomous systems.
IMPULSIVE DIFFERENTIAL EQUATION
Impulsive differential equations (IDEs) are differential equations that include
impulses, which represent sudden changes or perturbations in the system’s
behavior at specific points in time

Fixed Time IDEs


Incorporate impulses
that occur at
predetermined times

Variable Time IDEs


Incorporate impulses
that occur at variable
times, possibly
dependent on the state
of the system x(t)
MULTI AGENT SYSTEMS
Multi-agent systems (MASs) consist of
multiple interacting intelligent agents
that operate in a shared environment.
These agents are autonomous entities
capable of perceiving their environment,
processing information, making
decisions, and acting upon their
decisions. MAS find applications in
diverse domains, including artificial
intelligence, robotics, economics, social
sciences, and distributed computing.
TYPES OF MULTI AGENT SYSTEMS
Fixed
topology
Switching
topology

Directed
topology
Undirected
topology
CONCENSUS
Consensus refers to the process by Sampled-Data
which agents in a multi-agent Consensus
system reach an agreement on Quantized
certain shared quantities or states Consensus
through iterative information
Consensus in
exchange and updates. One of the
Random Networks
widely studied consensus
protocols is the linear consensus Scaled
protocol, where each agent Consensus
updates its state based on a linear Leader-Follower
combination of its own state and
Consensus
the states of its neighbors:
Finite-Time
Consensus
Bipartite
Consensus
FIXED TIME CONCENSUS
Fixed-time (FxT) consensus is the
concept of all agents reaching a
common value within a finite time
interval regardless of initial
conditions. The conditions of
every agent in a multi-agent
system that require them to reach
a common value by utilizing
communication with neighboring
agents within a finite amount of
time is known as the finite time
consensus problem.
LYAPUNOV’S DIRECT METHOD OF STABILITY
Lyapunov’s Direct Method is a powerful tool used in stability analysis of dynamical systems. As this method
provides stability information directly, without the need to solve the differential equations, we call it as the
”Lyapunov direct method”. The core of this approach revolves around the construction of a scalar function
denoted as W and its derivative W’. If certain properties are satisfied by W and its derivative W’ , then
several conclusions about the stability behaviour of the system can be drawn.
Consider the differential Equation:
(1)
. Assume that f(t, 0) = 0 for all t, then the following theorem holds true.:

(1)
LEADERLESS FIXED TIME CONSENSUS
(LLFXTC) OF MULTI AGENT SYSTEMS
Graph Theoritic Model
The communication between agents in a multi-agent system can be represented in the form of
a directed weighted graph G = (V, E, A), where V = {ν1, ν2, ν3...νN } denotes the set of
nodes in the graph, E ⊆ V × V is the set of edges representing the information of interaction
between agents, and A = [aij ] ∈ is the weighted adjacency matrix representing the edge
weights in the graph. . A directed edge ϵij ∈ E represents that there is a transmission of
information from the i th agent to the j th agent where the matrix entry aij represents the
weight of the edge eij .

Laplacian Matrix is defined as D-A, where D is the degree matrix. The second smallest
eigenvalue of the Laplacian matrix L, is denoted by (L) and is called the algebraic
connectivity.
Graph Theoritic Model (contd)
Following are some essential properties of the Laplacian of a graph.

Proof
MODEL DESCRIPTION
Consider a multi agent system consisting of N agents having the following general
nonlinear impulsive systems:
This is a function that describes
the continuous dynamics of the
system.It governs the behavior of
the system when there are no
impulsive changes.
(2)
This denotes the change in
This denotes the state of the state Z(t) at impulsive
the system at initial time times. The function G
depends on the state Z(t) and
the time t at these instances.

,
MODEL DESCRIPTION OF MAS
The origin of the system (2) is said to be globally Finite time stable (FTS) if it is
globally Lyapunov stable and there exists a function T: R n\0 → (0, ∞), such that
X(t, ) = 0 for all t ≥ T( , ). The function T is called the settling-time function.

The origin of system (2.1) is said to be globally FxTS (Fixed-time stable) if it is


globally FTS, and the settling-time function T is independent of

The inequality ensures that even though impulses may not occur exactly at regular
intervals, their number within any given time period (t,T) is predictable within certain
limits.
[18] - References
W belongs to the class φ, which typically includes functions that are continuous,
positive definite, and locally Lipschitz
Conditions for Global Finite-Time Stability

This inequality ensures that the


Lyapunov function W is bounded by
the state norm from both below and
above.W grows with the state norm
but is neither too small large, which
helps in proving stability.

This condition specifies that


At impulsive times, the Lyapunov
the time derivative of W is
function W after the impulse is
negative which indicates that
bounded by a factor ν times its value
W decreases over time,
before the impulse which shows that
leading the system towards
even after sudden changes (impulses),
stability.
the system maintains stability as W
does not increase excessively.
Settling-Time Bound
Given these conditions, the lemma states that the system is globally finite-time stable
(FxTS), meaning the system will reach stability within a finite amount of time.

It can be noted that the choice of the average impulsive interval τa > 0 and the positive
integer No have a strong influence on the stability dynamics, and these numbers might be
difficult to determine. So to overcome this constraint, the impulsive interval is bounded,
i.e., θ2 ≤ (tk+1 − tk) ≤ θ1, where θ1, θ2 are positive constants and measure T using θ1, θ2
Settling-Time Bound
An adaptation to the more practical scenario where impulses occur within bounded
intervals. The piecewise definition provides a more detailed and controlled description of
how the Lyapunov function W should behave to ensure finite-time stability under these new
conditions.
Using the new definition,
PROBLEM DESCRIPTION OF MAS

The control input governing the i-th agent is given by


PROBLEM DESCRIPTION OF MAS
The equations in matrix form can be written as

The multi agent system represented


by (2) under the effects of the control
protocol is said to achieve leaderless
fixed time consensus if the impulsive
interval follows , for some
positive constants θ1 and θ2. If
impulsive strength is such that
then an additional constraint has to
be satisifed in order to ensure
stability.
IMPULSIVE STRENGTH >1

With necessary Without necessary


Conditions Conditions
CONCLUSION
Less than 1

Beneficial for the


system stability The impulsive control leads to faster
convergence of the MAS, hence it is
beneficial for system stability. On the other
hand, stability in case of impulses with
Equal to 1 strength greater than one can be guaranteed
Impulsive if the sufficient condition is fulfilled. It is
The impulses have worthy to note that according to the derived
Strength no effect on the condition the average time interval between
system stability two impulses is directly proportional to the
impulsive strength and inversely proportional
to the coefficient of the linear term in the
Greater than 1 designed control protocol. This implies that
The impulse tends to for a given protocol, if the impulsive strength
cause disturbance in increases, the average time after which the
the system. next impulse is given must increase and this
also follows intuitively.
FUTURE ENDEAVOURS

My future endeavours include extending the results in case of Bipartite Concensus,


where the agents in a multi-agent system are divided into two disjoint groups or
partitions, and the objective is for the agents within each partition to reach
consensus, while the consensus values of the two partitions may differ

Later, in the next semester I would work upon extending these results for Leader
Follower Concensus. Leader-following MASs involve a group of agents where one
or more agents, known as leaders, influence the behavior of the rest of the agents,
known as followers, to achieve a common objective or goal.
REFERENCES
The references of the work can be found attached in the following

You might also like