ultimate numerical bound estimation of chaotic dynamical finance model
DESCRIPTION
This paper has investigated the boundedness of a 3D chaotic Dynamical Finance Model. Wehave discussed two bounds of this model. First by Lagrange multiplier method and second byoptimization method it was verified by using fmincon solver. Lyapunov Exponent calculatedusing Wolf algorithm and shown graphically in this paper. Lyapunov dimension of DynamicFinance Model also discussed. Numerical simulations are presented to show the effectiveness ofthe proposed scheme.TRANSCRIPT
Ultimate Numerical Bound Estimation of ChaoticDynamical Finance Model
Dharmendra Kumar1 Sachin Kumar2
1SGTB Khalsa CollegeDepartment of Mathematics
University of Delhi
2Department of MathematicsUniversity of Delhi
M3HPCST - 2015
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 1 / 17
Outline
1 Dynamical SystemReal Life Examples with chaotic system
2 Main Problem under discussion Ultimate Bound
3 Four Methods to estimate ultimate bound of chaotic systems.
4 Dynamic Finance ModelDynamical behaviour of the systemPhase diagrams for the Dynamic Finance ModelLyapunov Exponent for Dynamic Finance ModelLyapunov Dimension
5 Langrange Multiplier Method
6 Optimization MethodMain Optimization ProblemNumerical Calcualtions for BoundEllipsoidal Bound
7 Future Work
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 2 / 17
Dynamical System
Dynamical System
Dynamics is the study of change
A Dynamical System is just a recipe for saying how a system ofvariables interacts and changes with time.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 3 / 17
Dynamical System
Dynamical System
Dynamics is the study of change
A Dynamical System is just a recipe for saying how a system ofvariables interacts and changes with time.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 3 / 17
Dynamical System Real Life Examples with chaotic system
Real Life Examples with chaotic system
Some real life examples are as follows:
The solar system (Poincare in 1888)
Complex behaviour in Hamiltonian mechanics(Birkhoff, Kolmogorov, Arnold & Moser in 1920 - 1960)
The Weather Forcasting (Lorenz in 1963)
Turbulence in fluids (Libchaber)
Solar activity (Parker)
Population growth or Logistic Map(May in 1970)
lots and lots of other systems
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 4 / 17
Dynamical System Real Life Examples with chaotic system
Real Life Examples with chaotic system
Some real life examples are as follows:
The solar system (Poincare in 1888)
Complex behaviour in Hamiltonian mechanics(Birkhoff, Kolmogorov, Arnold & Moser in 1920 - 1960)
The Weather Forcasting (Lorenz in 1963)
Turbulence in fluids (Libchaber)
Solar activity (Parker)
Population growth or Logistic Map(May in 1970)
lots and lots of other systems
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 4 / 17
Dynamical System Real Life Examples with chaotic system
Real Life Examples with chaotic system
Some real life examples are as follows:
The solar system (Poincare in 1888)
Complex behaviour in Hamiltonian mechanics(Birkhoff, Kolmogorov, Arnold & Moser in 1920 - 1960)
The Weather Forcasting (Lorenz in 1963)
Turbulence in fluids (Libchaber)
Solar activity (Parker)
Population growth or Logistic Map(May in 1970)
lots and lots of other systems
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 4 / 17
Dynamical System Real Life Examples with chaotic system
Real Life Examples with chaotic system
Some real life examples are as follows:
The solar system (Poincare in 1888)
Complex behaviour in Hamiltonian mechanics(Birkhoff, Kolmogorov, Arnold & Moser in 1920 - 1960)
The Weather Forcasting (Lorenz in 1963)
Turbulence in fluids (Libchaber)
Solar activity (Parker)
Population growth or Logistic Map(May in 1970)
lots and lots of other systems
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 4 / 17
Dynamical System Real Life Examples with chaotic system
Real Life Examples with chaotic system
Some real life examples are as follows:
The solar system (Poincare in 1888)
Complex behaviour in Hamiltonian mechanics(Birkhoff, Kolmogorov, Arnold & Moser in 1920 - 1960)
The Weather Forcasting (Lorenz in 1963)
Turbulence in fluids (Libchaber)
Solar activity (Parker)
Population growth or Logistic Map(May in 1970)
lots and lots of other systems
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 4 / 17
Dynamical System Real Life Examples with chaotic system
Real Life Examples with chaotic system
Some real life examples are as follows:
The solar system (Poincare in 1888)
Complex behaviour in Hamiltonian mechanics(Birkhoff, Kolmogorov, Arnold & Moser in 1920 - 1960)
The Weather Forcasting (Lorenz in 1963)
Turbulence in fluids (Libchaber)
Solar activity (Parker)
Population growth or Logistic Map(May in 1970)
lots and lots of other systems
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 4 / 17
Main Problem under discussion Ultimate Bound
Main Problem
Ultimate Bound
The estimate of the ultimate bound for a chaotic system is of greatimportance for chaos control, chaos synchronization, Hausdorff dimensionand the Lyapunov dimension of chaotic attractors
There are a few studies which have discussed the solution bounds andinvariant sets of the chaotic systems, such as
The Lorenz system in 2001
The hyperchaotic Lorenz–Haken system in 2009
Lorenz–Stenflo system in 2015
Stochastic cellular neural networks with delays in 2011
One tumor growth model in 2013
A class of HDQA chaotic systems in 2011
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 5 / 17
Four Methods to estimate ultimate bound of chaotic systems.
How to estimate the bound of strange attractor?
Task is difficult
Chaotic systems are bounded, the bounds have important application inchaos control and synchronization; But it is a difficult task to estimate theultimate bounds of them.
There are mainly four methods to estimate the bounds of chaotic systemsin current literatures. Four popular methods are following:
The hyper plane oriented method
The iteration theorem and the first order extremum theorem,
Lyapunov stability theory combined with the comparison principlemethod
The optimization method.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 6 / 17
Dynamic Finance Model Dynamical behaviour of the system
Dynamic Finance Model
Ma JH & Chen YS in 2002 reported a dynamic model of finance,composed of three first-order differential equations. The model describesthe time-variation of three state variables
x = z + (y − a)x
y = 1− by − x2 (1)
z = −x − cz
x: interest rate.
y: investment demand.
z: price index.
where a is the saving amount, b is the cost per investment, c is theelasticity of demand. Dynamical behaviour of the system:
Symmetry and Invariance
Dissipativity and Existence of Attractor
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 7 / 17
Dynamic Finance Model Phase diagrams for the Dynamic Finance Model
Phase diagrams for the Dynamic Finance Model.
−2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.51.5
2
2.5
3
3.5
4
4.5
5
5.5
x
y
−2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5−1.5
−1
−0.5
0
0.5
1
1.5
x
z
1.5 2 2.5 3 3.5 4 4.5 5 5.5−1.5
−1
−0.5
0
0.5
1
1.5
y
z
−2
−1
0
1
2
0
2
4
6
−3
−2
−1
0
1
2
3
zy
x
Figure: Phase diagrams for the Dynamic Finance Model
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 8 / 17
Dynamic Finance Model Lyapunov Exponent for Dynamic Finance Model
Lyapunov Exponent
Numerical Values of Lyapunov Exponent
L1 = 0.7848, L2 = 0.2260, L3 = −1.3332
Why Finance Model is chaotic?
Two L1, L2 are positive Lyapunov exponent, and the third one is negative.Thus, the system is chaotic. The time histories, phase diagrams, and thelargest Lyapunov Exponent were used to identify the dynamics of thesystem. The largest Lyapunov Exponent were calculated using the schemeproposed by Wolf in 1985.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 9 / 17
Dynamic Finance Model Lyapunov Dimension
Lyapunov Dimension for finance chaotic system
The Lyapunov dimension of the chaotic system (1) is given by
DL = j +1
|Lj+1|
j∑i=1
Li = 2 +L1 + L2|L3|
= 2.7582 (2)
0 200 400 600 800 1000−2
−1.5
−1
−0.5
0
0.5
1Dynamics of Lyapunov exponents
Time
Lya
pu
no
v e
xp
on
en
ts
Figure: Lyapunov Exponent of the Dynamic Finance Model.
So, the chaos in the system is very obvious.Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 10 / 17
Langrange Multiplier Method
Langrange Multiplier Method
Consider the following problem which is constructed using Lyapunovfunction theory
Max V (x , y , z) =1
2(x2 + y2 + (z − a)2) (3)
Subject to constraint:
Γ : 3(x − 1
2)2 +
1
10(y − 5)2 + (z − 3)2 =
11
4(4)
Estimate of ultimate bound for (1) is
Ω =
(x , y , z)|1
2(x2 + y2 + (z − 3)2) ≤ 52.5971
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 11 / 17
Optimization Method
Optimization Method using matrix analysis
Theorem
Suppose that there exits a real symmetric matrix P > 0 and a vectorµ ∈ R3 such that
Q = ATP + PA + 2(BT1 PuT ,BT
2 PuT , ...,BTn PuT )T < 0
and∑
xiXT (BT
i P + PBi )X = 0 for any X = (x1, x2, ..., xn)T ∈ R3 andu = (u1, u2, ..., un) = 2µTP then (1) is bounded and has the followingultimate bound set also called positively invarient set:
Ω =X ∈ R3|(X + µ)TP(X + µ) ≤ Rmax
(5)
Rmax to be calculated.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 12 / 17
Optimization Method Main Optimization Problem
Main Optimization Problem
Rmax is a real number to be determined by following OptimizationProblem:Objective Function:
Max (X + µ)TP(X + µ) (6)
subject to
XTQX + 2(µTPA + CTP)X + 2CTPµ = 0. (7)
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 13 / 17
Optimization Method Numerical Calcualtions for Bound
Numerical Calcualtions for Bound
Bounds are numerically calculated and observed for different parameter. Inthis paper, we have shown the ellipsoidal boundedness of the Financialdynamic Model.
Parameters Values Initial Condition (x0, y0, z0) Rmax
p11 = 1, p33 = 2 (2, 3, 2) 281560.49
p11 = 0.1, p33 = 0.2 (2, 3, 2) 129247.49
p11 = 0.001, p33 = 0.001 (2, 3, 2) 56796
p11 = 0.001, p33 = 0.002 (2, 3, 2) 79519.34
p11 = 1000, p33 = 2 (2, 3, 2) 1.2E48
p11 = 1, p33 = 2000 (2, 3, 2) 1.2E12
p11 = 0.001, p33 = 0.001 (2, 3, 2) 56796
p11 = 1, p33 = 2 (0, 0, 0) 165775.15
p11 = 1, p33 = 1 (2, 3, 2) 639152.0
Table: Simulation results using MATLAB
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 14 / 17
Optimization Method Ellipsoidal Bound
Ellipsoidal Bound
Attractor formed using finance chaotic system is enclosed with an ellipsoidwhich is shown in the following fig:
−4
−2
0
2
4
0
2
4
6−4
−2
0
2
4
6
x(t)y(t)
z(t
)
−10
−5
0
5
10
−10
−5
0
5
10−5
0
5
10
15
y(t)z(t)
x(t
)
Figure: The bound of the chaotic attractor of system with a = 3, b = 0.1 andc = 1, p11 = 1, p33 = 2.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 15 / 17
Future Work
Future Work
New Problems in this topic
At present, ultimate bound estimations of many chaotic systems are still adifficult mathematical problem, such as the well-known
Chen system,
Lu system and
multi-scroll chaotic attractors.
How to find explicit ultimate bound sets for these systems is still anunsolved question.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 16 / 17
Summary
For Further Reading I
S.H. Strogatz.Nonlinear Dynamics And Chaos.Studies in Nonlinearity, 2000.
Wang P, Li D, Wu X, L J, Yu X.Ultimate bound estimation of a class of high dimensional quadraticautonomous dynamical systems.Int. J. Bifurc. Chaos Appl. Sci. Eng., 21(9):2679–2694, 2011.
Leonov GA.Bound for attractors and the existence of homoclinic orbit in Lorenzsystem.J. Appl. Math. Mech., 65:19–32, 2001.
Dharmendra Kumar, Sachin Kumar Ultimate Numerical Bound Estimation of Chaotic Dynamical Finance ModelM3HPCST - 2015 17 / 17