Scribd
Upload
66 views8 pages

Analyzing Nullclines in Phase Planes

This document provides an overview of using nullclines to sketch and analyze phase planes for systems of differential equations. It defines x-nullclines and y-nullclines as sets of points where the direction of flow is horizontal or vertical, found by setting expressions involving f(x,y) and g(x,y) equal to zero. The document uses an example system to demonstrate finding the nullclines algebraically, sketching the direction field, and identifying equilibrium points where the nullclines intersect. It also summarizes solving homogeneous linear systems of differential equations by reducing to a second-order equation and relating its characteristic roots to eigenvalues.

Uploaded by

RizkyW
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
66 views8 pages

Analyzing Nullclines in Phase Planes

This document provides an overview of using nullclines to sketch and analyze phase planes for systems of differential equations. It defines x-nullclines and y-nullclines as sets of points where the direction of flow is horizontal or vertical, found by setting expressions involving f(x,y) and g(x,y) equal to zero. The document uses an example system to demonstrate finding the nullclines algebraically, sketching the direction field, and identifying equilibrium points where the nullclines intersect. It also summarizes solving homogeneous linear systems of differential equations by reducing to a second-order equation and relating its characteristic roots to eigenvalues.

Uploaded by

RizkyW
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

Math 221 - Section 5

A quick guide to sketching phase planes


Our text discusses equilibrium points and analysis of the phase plane. However, there is one
idea, not mentioned in the book, that is very useful to sketching and analyzing phase planes, namely
nullclines. Recall the basic setup for an autonomous system of two DEs:
dx
= f (x, y)
dt
dy
= g(x, y)
dt
To sketch the phase plane of such a system, at each point (x0 , y0) in the xy-plane, we draw a vector
starting at (x0 , y0 ) in the direction f (x0 , y0)i + g(x0 , y0)j.
dx
Definition of nullcline. The x-nullcline is a set of points in the phase plane so that = 0.
dt
Geometrically, these are the points where the vectors are either straight up or straight down. Alge-
braically, we find the x-nullcline by solving f (x, y) = 0.
dy
The y-nullcline is a set of points in the phase plane so that = 0. Geometrically, these are the
dt
points where the vectors are horizontal, going either to the left or to the right. Algebraically, we
find the y-nullcline by solving g(x, y) = 0.
How to use nullclines. Consider the system
dx  x
= 2x 1 − − xy,
dt 2
dy  y
= 3y 1 − − 2xy.
dt 3
 x
To find the x-nullcline, we solve 2x 1 − − xy = 0, where multiplying out and collecting the
2
common factor of x gives x(2 − x − y) = 0. This gives two x-nullclines, the line x + y = 2 and the
y-axis. Solutions of this system move to the right if x(2 − x − y) > 0, so the direction field arrows
will point toward the right (and either upward or downward) in two cases: When both x > 0 and
(2 − x − y) > 0 (to the right of the y axis and below the line y = −x + 2), and when both x < 0
and (2 − x − y) < 0 (to the left of the y axis and above the line y = −x + 2).
y
3

1 2 3 x
 y
To find the y-nullcline, we solve 3y 1 − − 2xy = 0, where multiplying out and collecting the
3
common factor of y gives y(3 − y − 2x) = 0. This gives two y-nullclines, the line 2x + y = 3
and the x-axis. Solutions of this system move upward if y(3 − y − 2x) > 0, so the direction field
arrows will point up (and either right or left) in two cases: When both y > 0 and (3 − y − 2x) > 0
(above the x axis and below the line y = −2x + 3), and when both y < 0 and (3 − y − 2x) < 0
(below the x axis and above the line y = −2x + 3).
y
3

1 2 3 x

Combining this information gives us the following picture. Notice that we can draw directions
on each nullcline by using the direction information from the other graph. For example, the line
segment from (1, 1) to (0, 3), since it is above the line y = −x + 2 and to the right of the y-axis,
has solutions moving to the left.

y
3

x
1 2 3

dx dy
Also, where the x-nullcline and y-nullcline cross, both and are zero. So these points
dt dt
(marked by dots in the above graph) are equilibrium points.
Once a solution enters the triangle with vertices (1, 1), (0, 2) and (0, 3), it can never leave.
Similarly, solutions in the triangle with vertices (1, 1), (3/2, 0) and (2, 0) can never leave.
Exercises. Graph the nullclines, sketch the direction fields, and discuss the possible fates of solu-
tions for the following systems. Note: The nullclines may not be straight lines.
dx dy
(1) = x(−x − 3y + 150), = y(−2x − y + 100).
dt dt
dx dy
(2) = x(10 − x − y), = y(30 − 2x − y).
dt dt  
dx  x dy 9
(3) = 2x 1 − − xy, =y − y − x2 y.
2
dt 2 dt 4
dx dy
(4) = x(−4x − y + 160), = y(−x2 − y 2 + 2500).
dt dt

2
Math 221 - Section 5
A quick guide to homogeneous linear systems of DEs
A standard form homogeneous linear system of two first order constant coefficient differential
equations is
dx dy
= ax + by = cx + dy
dt dt
where a, b, c, and d are constants. Both x and y are dependent variables, depending on the
independent variable t. The solution to this system is a pair of functions x(t) and y(t) that satisfy
both DEs together. Often, it will be convenient to write the solution as a column vector
 
x(t)
~s(t) = .
y(t)

To solve this system when b 6= 0, we use the first DE to eliminate the variable y from the second
DE. Solving the first DE for y:
1
y = (x0 − ax) .
b
Then the derivative
1 a
y 0 = x00 − x0 .
b b
Plugging these into the second DE of the system gives
1 00 a 0 1
x − x = cx + d( (x0 − ax)) .
b b b
Rearranging the terms of this equation and multiplying both sides by b, we now have one second
order linear constant coefficient DE to solve for x, namely
x00 − (a + d)x0 + (ad − bc)x = 0 .
So we solve this DE in the usual way (chap. 3 sec. 3 of the text). Once we have x, we can use the
formula in the first box to find y.
Using the vocabulary discussed in chapter 5 of the text, the characteristic roots r1 , r2 from the
auxiliary equation for the constant coefficient second order homogeneous DE above for x are also
the eigenvalues of the matrix  
a b
.
c d
 
1
An eigenvector associated to the root ri is the vector ri −a . These vectors are discussed further
b
in examples below.
Homogeneous linear systems always have a critical point at (0, 0); that is, the constant functions
x(t) ≡ 0 and y(t) ≡ 0 are solutions of this system for any values of a, b, c, and d. The behavior
around the critical point (0,0) depends on the characteristic roots of the auxiliary equation. For
example, the critical point can be stable, unstable (a source), or asymptotically stable (a sink).
The critical point is described in various ways as a node, saddle, spiral, center, or star. See the
pictures on the last page of this handout for pictorial definitions of these terms.
Example (A): Find the general solution to the linear system
dx dy
= 0x − 2y , = x + 3y ,
dt dt
and analyze the phase plane direction field for this system.
Answer: Here we have b = −2 6= 0 in the first DE, we can solve for y and get y = − 12 x0 . Plugging
this into the second DE, then − 12 x00 = x + 3(− 21 x0 ). Then the associated second order equation
1
(also from the boxed formula on p. 1) is x00 − 3x0 + 2x = 0. This has roots 1, 2, so the solution for
x is
x(t) = C1 et + C2 e2t .
Using the formula (above or from the boxes) for y(t), we get y = − 12 x0 = − 12 C1 et − C2 e2t .
The general solution for the system is
1
x(t) = C1 et + C2 e2t y(t) = − C1 et − C2 e2t .
2
     
x(t) 1 1
Written as a vector, the general solution is ~s(t) = = C1 et + C2 e2t
.
y(t) − 12 −1
(Note : It is important to notice that the same constants appear in the solutions for both x(t) and
y(t), so we cannot replace a multiple of an arbitrary constant with an arbitrary constant in only
one of the solutions.)
The x-nullcline is the line y = 0; arrows point to the right when y < 0 and to the left when
y > 0. The y-nullcline is the line y = − 13 x; arrows point upward when y > − 31 x and downward
when y < − 31 x.
The solution with C1 = 1 and C2 = 0 is given by x(t) = et , y(t) = − 12 et ; then for all t, this
solution follows the ray y = − 21 x for
 x > 0 and y < 0 in the phase plane, going up and to the right
1
in the direction of the eigenvector for the eigenvalue/root r1 = 1 as t increases. The solution
− 12
with C1 = 0 and C2 = 1 has x(t) = e2t , y(t) 2t
 = −e , and so follows the ray y = −x with x > 0 and
1
y < 0 in the direction of the eigenvector for the eigenvalue/root r2 = 2 as t → ∞.
−1
For the solution of the IVP with x(0) = 12 and y(0) = −5, as t → ∞ we have x(t) → −∞ and
y(t) → ∞. However, for this IVP the solution (x(t), y(t)) can never reach the point (−7, 6).
The critical point (0, 0) is unstable; the characteristic roots 1,2 for x distinct positive real num-
bers, and so the critical point is a node source.

Example (B): Find the general solution of the system


dx dy
= x + 2y , = −5x − y ,
dt dt
and analyze stability of the critical point (0,0).
2
Answer: We have b = 2 6= 0 in the first DE, we can solve for y and get y = 21 x0 − 21 x. Plugging
this into the second DE, then 12 x00 − 21 x0 = −5x − ( 12 x0 − 21 x). Then the associated second order
equation is x00 + 9x = 0. This has roots ±3i, so the solution for x is
x(t) = C1 sin(3t) + C2 cos(3t).
Using the formula for y(t), we get
y(t) = (1/2)(3C1 cos(3t) − 3C2 sin(3t)) − (1/2)(C1 sin(3t) + C2 cos(3t))
= C1 ((3/2) cos(3t) − (1/2) sin(3t)) + C2 (−(1/2) cos(3t) − (3/2) sin(3t)) .
Written as a vector, the general solution is
     
x(t) sin(3t) cos(3t)
~s(t) = = C1 + C2 .
y(t) (3/2) cos(3t) − (1/2) sin(3t) −(1/2) cos(3t) − (3/2) sin(3t)
This general solution involves only cosine and sine functions, but not exponential functions, so
x(t) and y(t) can’t (asymptotically) approach 0, and they also can’t go to ±∞. The critical point
(0,0) for this linear system is a stable center.
Example (C): Two large tanks are interconnected by pipes. Fresh water flows into tank A at a
rate of 6 L/min. Solution flows through a pipe from tank A to tank B at a rate of 8 L/min, and
solution flows through another pipe from tank B to tank A at a rate of 2 L/min. The solution in
tank B also drains out onto the ground at a rate of 6 L/min. The solutions in both tanks are kept
evenly mixed. Initially tank A contains 3 g of salt dissolved in 24 L of water, and tank B contains
5 g of salt dissolved in 24 L of water. Set up an initial value problem modeling this situation.
Answer: Let t be the time in minutes, with initial time t = 0. Let x(t) be the mass of salt in
grams in tank A at time t, and let y(t) be the mass of salt in grams in tank B at time t. The initial
conditions say that x(0) = 3 and y(0) = 5.
The volume of solution in tank A is changing at a rate of 6 − 8 + 2 = 0 L/min, so the volume in
that tank is a constant 24 L. The volume in tank B changes at a rate of 8 − 2 − 6 = 0 L/min, so
the volume in that tank is also a constant 24 L/min.
First analyze the input and output of salt at tank A. The fresh water flowing into tank A contains
0 g/L of water, so the input from this pipe changes the mass of salt in tank A at a rate of (0 g/L)(6
L/min)=0 g/min. The solution flowing from tank A to tank B removes salt from tank A at a rate
of ( x(t)
24 g/L)(8 L/min). The solution flowing from tank B to tank A inputs salt into tank A at a
rate of ( y(t) dx dx 2y 8x
24 g/L)(2 L/min). Since dt = (input rate) - (output rate), then dt = 0 + 24 − 24 g/min.
Next do the same for tank B: dy 8x 2y 6y
dt = (input rate) - (output rate) = 24 − 24 − 24 g/min.
The initial value problem modeling this situation is:
dx 1 1 dy 1 1
=− x+ y , = x − y ; x(0) = 3 , y(0) = 5 .
dt 3 12 dt 3 3
Example (D): Solve the initial value problem from Example (C), and discuss what happens to
the solution as t → ∞.
Answer. Here we have a = − 13 , b = 121
, c = 31 , and d = − 31 , with b 6= 0. The associated second
1 1 0 1 1 1
00
order DE is x − (− 3 − 3 )x + ((− 3 )(− 3 ) − ( 12 )( 13 ))x = 0, or x00 + 23 x0 + 12
1
x = 0. This DE has
2 2 1 1 1
auxiliary equation r + 3 r + 12 = 0, with roots r1 = − 6 and r2 = − 2 . Then the solution for x is
x(t) = C1 e−(1/6)t + C2 e−(1/2)t .
Then
1
y = (1/ 12 )(x0 − (− 13 )x) = 12x0 + 4x
1
= 12(− 6 C1 e−(1/6)t − 12 C2 e−(1/2)t ) + 4(C1 e−(1/6)t + C2 e−(1/2)t )
= 2C1 e−(1/6)t − 2C2 e−(1/2)t
3
Plugging the initial values into this general solution gives
x(0) = C1 + C2 = 3 y(0) = 2C1 − 2C2 = 5.
These equations have solution C1 = 11 1
4 and C2 = 4 . Then the solution of the initial value problem
is
11 1 11 1
x(t) = e−(1/6)t + e−(1/2)t , y(t) = e−(1/6)t − e−(1/2)t .
4 4 2 2
As t → ∞, the functions x(t) and y(t) in these solutions (asymptotically) approach 0. (The
critical point (0,0) for this linear system is a node sink.)

Exercises:
(1) Find the general solution to the linear system if b is zero. Hint: The first DE has no y’s, so
solve it first and plug that solution into the second DE.
For exercises (2)-(6), find the general solution (or IVP solution when initial values are given),
draw the phase plane direction field, and discuss the stability of the critical point (0,0):
dx 1 dy
(2)=(4.1.15) = y, = −8x .
dt 2 dt
dx dy
(3)=(4.2.3) = −3x + 2y , = −3x + 4y ; x(0) = 0 , y(0) = 2 .
dt dt
dx dy
(4)=(4.2.6) = x + 9y , = −2x − 5y ; x(0) = 3 , y(0) = 2 .
dt dt
dx dy
(5)=(6.1.18) = −y , = 4x .
dt dt
dx dy
(6)=(6.1.20) =y , = −5x − 4y .
dt dt
Exercises (7)-(8) deal with a two-tank system. Fresh water flows into tank A, and brine flows
from tank B onto the ground, both at a rate of 2r L/min. Brine flows from tank B to tank A at a
rate of r L/min, and brine flows from tank A to tank B at a rate of 3r L/min. The initial amounts
x(0) and y(0) of salt in tanks A and B, respectively, are given, as are the volumes VA and VB in
the tanks. Set up and solve the initial value problem modeling the situation in each exercise.
(7)=(4.2.30) r = 10 L/min x(0) = 0.5 g, y(0) = 0.5 g, VA = 100 L, VB = 200 L.
(8) r = 3 L/min x(0) = 1 g, y(0) = 2 g, VA = 20 L, VB = 20 L.

4
Math 221 - Section 5
A quick guide to homogeneous linear systems of DEs and the phase plane
A standard form homogeneous linear system of two first order differential equations is
dx dy
= ax + by = cx + dy
dt dt
where a, b, c, and d are constants.
Homogeneous systems always have a critical point at (0, 0); that is, the constant functions
x(t) ≡ 0 and y(t) ≡ 0 are solutions of this system for any values of a, b, c, and d.
To solve the system when b 6= 0, we solve the first DE for y and using this to eliminate
the variable y from the second DE:
1
y = (x0 − ax) x00 − (a + d)x0 + (ad − bc)x = 0 .
b
The behavior around the critical point (0,0) depends on the characteristic roots of the aux-
iliary equation. For example, the critical point can be stable, unstable (a source), or asymp-
totically stable (a sink). The critical point is described in various ways as a node, saddle,
spiral, center, or star. See the pictures on the last page of this handout for pictorial definitions
of these terms. Exercise (1): In Example (A) of the Solving linear systems handout we
dy
found that the general solution to the linear system in dx
dt
= x+3y , dt
= 0x+2y
is
x(t) = 3C1 e2t + C2 et y(t) = C1 e2t .
Graph the two solutions where C1 = 0 and C2 = 1 and where C1 = −1 and C2 = 0 in
the phase plane. What does this tell you about (0, 0)? Example (B): In Example (B) of
the Solving linear systems handout we found that the general solution to the linear system
dx dy
dt
= x + 2y , dt
= −5x − y
is  
x(t) = C1 sin(3t)+C2 cos(3t) , y(t) = C1 32 cos(3t) − 21 sin(3t) +C2 − 21 cos(3t) − 32 sin(3t) .
This general solution involves only cosine and sine functions, but not exponential functions,
so x(t) and y(t) can’t (asymptotically) approach 0, and they also can’t go to ±∞. The critical
point (0,0) for this linear system is a stable center. Example (C-D): In Examples (C-D) of
the Solving linear systems handout we found that the general solution to the linear system
dx dy
dt
= − 31 x + 12
1
y, dt
= 31 x − 13 y
is
x(t) = C1 e−1/6)t + C2 e−(1/2)t y(t) = 2C1 e−(1/6)t − 2C2 e−(1/2)t .
As t → ∞, the functions x(t) and y(t) in these solutions (asymptotically) approach 0.
The critical point (0,0) for this linear system is a node sink. Exercises: For the following
systems, find the general solution and discuss the stability of the critical point (0, 0): (2)
dx dy dx dy
= 4x , = x + 4y . (3) = 4x − 5y , = 2x + 6y .
dt dt dt dt
distinct real roots, both positive distinct real roots, both negative distinct real roots, different signs

4 4 4
y y y
2 2 2

–4 –2 0 2 x 4 0 –4 –2 0 2 x 4
–4 –2 2 x 4

–2 –2 –2

–4 –4 –4

node source (unstable) node sink (asymptotically stable) saddle point (unstable)

complex roots, real part positive complex roots, real part negative complex roots, real part zero

4 4 4
y y y
2 2 2

–4 –2 0 2 x 4 –4 –2 0 2 x 4 –4 –2 0 2 x 4

–2 –2 –2

–4 –4 –4

spiral source (unstable) spiral sink (asymptotically stable) center (stable)

repeated roots, positive repeated roots, negative repeated roots, diagonal matrix

4 4 4
y y y
2 2 2

–4 –2 0 2 x 4 –4 –2 0 2 x 4 –4 –2 0 2 x 4

–2 –2 –2

–4 –4 –4

improper node, source (unstable) improper node, sink (asymptotically stable) star point

Common questions

Powered by AI

The characteristic roots of the auxiliary equation can be used to determine the stability of the critical point (0, 0) in a homogeneous linear system of differential equations. These roots are obtained from the characteristic equation derived from the second-order linear constant coefficient differential equation associated with the system. The nature of the roots dictates the system's behavior around the critical point. If both roots are real and positive, the critical point is a node source and unstable. If they are real and negative, it is a node sink and asymptotically stable. For complex roots, if the real part is negative, the system describes a spiral sink and is asymptotically stable; if positive, it is a spiral source and unstable. Real roots of opposite signs indicate a saddle and instability .

In the analysis of homogeneous linear systems of differential equations, eigenvalues and eigenvectors play a crucial role in determining the phase plane behavior. The eigenvalues, derived from the auxiliary equation, indicate the type and stability of the critical points. For example, distinct positive real eigenvalues indicate an unstable node, while distinct negative real values indicate a stable node. Complex eigenvalues with negative real parts describe a spiral sink, suggesting asymptotic stability, while positive real parts result in a spiral source, indicating instability. Eigenvectors correspond to the directions of trajectories in the phase space and indicate the principal directions along which solutions tend to grow or decay. These vectors directly influence the trajectory paths in the phase plane, thereby guiding the overall system dynamics .

When solving a standard form homogeneous linear system of two first-order differential equations with b ≠ 0, the first step is to express y in terms of x and its derivatives using the first differential equation. Specifically, solve for y as y = (1/b)(x' - ax). Then, substitute this expression for y into the second differential equation to eliminate y and obtain a single second-order linear differential equation in terms of x. Solve this equation to find x(t). After determining x(t), substitute back into the relationship to obtain y(t). This yields the solutions for x(t) and y(t), which can be expressed as a column vector .

The interaction between two large interconnected tanks with constant flow rates can be modeled using a system of differential equations by defining variables for the amounts of a substance in each tank, typically as functions of time. Each rate of change of the substance in a tank is described by equations governing inflow and outflow rates. For instance, if tank A receives fresh flow and has outflow to tank B, and vice versa, differential equations can represent changes in concentration or amount over time, considering these flows. The differential system thereby captures the dynamic behavior, with initial conditions specifying starting quantities. This model informs about long-term behavior by solving the system, determining equilibrium points, and examining stability through eigenvalues. Stable systems converge to equilibrium, while unstable ones diverge, with flow rates determining if the tanks reach a consistent distribution of substance or fluctuate indefinitely .

Real eigenvalues in a system's matrix representation lead to phase plane features like nodes and saddles, indicating solutions that either converge directly to or diverge from equilibrium (stability or instability). If eigenvalues are distinct and real, the phase plane reflects straightforward approaches (nodes) or repulsions (saddle points). In contrast, complex eigenvalues indicate rotational behavior, leading to spiral sinks or sources, reflecting solutions that spiral towards or away from the equilibrium point. The presence of purely complex eigenvalues results in centers, orbits that loop endlessly around the critical point without convergence or divergence, signifying neutral stability in the plane .

Introducing differential flow rates in a system's differential equations modifies the balance of inflow and outflow terms, impacting the rates of change for each tank's substance concentration. Such differential rates introduce non-symmetry in the equations, leading to changes in equilibrium points and potentially altering the stability analysis. This can shift the system from stability to instability or vice versa, depending on whether the flow rates increase the net substance concentration in a tank or deplete it unevenly. Stability is assessed by solving the modified system to find eigenvalues, which define if the system moves toward equilibrium or diverges over time. Unequal rates can introduce more complex behaviors and necessitate additional considerations in solving the differential equations to predict long-term system dynamics .

Nullclines in the context of phase plane analysis are the sets of points in the phase plane where one of the derivatives of the system becomes zero. Specifically, the x-nullcline consists of points where dx/dt = 0, meaning the vectors are straight up or down. In contrast, the y-nullcline consists of points where dy/dt = 0, meaning the vectors are horizontal, pointing left or right. They help in sketching the phase plane by indicating where the direction of the vector field changes. By solving f(x, y) = 0 and g(x, y) = 0 algebraically, we find these nullclines. The intersections of x- and y-nullclines identify equilibrium points. This information aids in visualizing the trajectory of solutions and the behavior of the system within different regions demarcated by the nullclines .

When b = 0 and c ≠ 0 in a linear system defined by dx/dt = ax + by and dy/dt = cx + dy, the solution approach changes due to the inability to automatically eliminate y from the second equation as when b ≠ 0. Instead, you solve the first equation dx/dt = ax directly as it involves only x and its derivatives. Upon solving for x(t), you then substitute it into the second equation to find a differential equation in terms of y that includes x's solution. Solving this equation gives y(t). This method involves first resolving the simpler decoupled equation, which does not depend on y directly, then solving the remaining equation influenced by the first solution .

Yes, a homogeneous linear system of differential equations can have a critical point that is neither asymptotically stable nor unstable. This occurs under conditions where the critical point is a center. A center is characterized by purely imaginary eigenvalues, indicating the existence of oscillatory behavior without exponential growth or decay. On the phase plane, this is visualized as closed orbits or ellipses around the critical point, with trajectories presenting periodic solutions. Despite being stable (as solutions neither grow unbounded nor decay to zero), the critical point's neutral stability means it does not attract or repel solutions asymptotically .

The behavior of a system within a defined region of the phase plane, such as a triangle bounded by nullclines, is determined by the directions dictated by the vector field in each segment, as informed by the nullclines. When nullclines intersect to form a region, they set boundaries for solution trajectories. Solutions cannot cross these lines unless at intersections, meaning behavior within is dictated by nullcline boundaries and direction fields. For example, solutions may never leave a triangle formed by nullclines due to consistent inward pointing vectors. This confinement amplifies the role of steady-state analysis, as stability or instability within this region determines the fate of trajectories, such as converging to an equilibrium or perpetually circulating .

You might also like