Sie sind auf Seite 1von 5

MTH256 Lab 2

Justin Drawbert draw@pdx.edu March 14, 2012


In this lab we consider a predator prey model which can be represented by the non-linear system of dierential equations dx = 9x x2 3xy dt dy = 2y + xy dt where 0. Here x represents the population of prey (in some scaled units), and y represents the population of predators1 (likewise, in scaled units). In particular, we will investigate what happens to these populations as t for dierent values of . We will examine this in more detail, but to get a quick sense of what is going on here, we take a look at each of the terms. We may think of the 9x term as the growth rate of the prey. Then, we may make sense of the x2 term by considering y = 0 (a system with no predators). Here, dx/dt will take the form of the logistic function dx = 9x x2 = x(9 x), dt which we recall is the form of the logistic equation, with carrying capacity x = 9/. With y again, nonzero, the 3xy describes the tendency of the predator to consume more when more prey is available. Then considering dy/dt with x = 0 we see the 2y term dictates that our predator cannot survive without the prey on which it feeds and will eventually vanish. The xy term basically says2 that the rate of population growth for our predator will be proportionate to both x and y.

Equilibrium Points
A valid question here might be where the equilibrium points of the system might be. To get a better idea of whats going on here, we rst compute the equilibrium points in a similar manner as we have done before. That is, by taking the right hand side of both equations, setting each to 0 and solving for x and y. This gives the system 9x x2 3xy 2y + xy =0 =0

Solutions become a bit more apparent if we express this as x (9 x 3y) = 0 y (2 + x) =0


1 In dierential equations texts as well as in lectures an oft used example is that of foxes as predators and rabbits as prey. The author acknowledges that, to a compassionate reader with a vivid imagination, this might be a bit of a gruesome and thus undesirable example. So rather than think of cute, innocent bunnies being devoured by razor toothed foxes, one might consider instead the prey x as being the kale plants in the authors garden. Then y is merely a representation of the authors uninhibited desire to eat kale. 2 Or with the kale analogy, this says the more kale the author eats, the greater the increase in the authors desire to eat kale.

The trivial solution


Right o the bat (0, 0) is a solution, but given the nature of the problem at hand, this is trivial. All this says is that if both our predators and prey are extinct, they are just that, extinct. Fair enough. Moving along, then. . .

No predators
Another solution that might be easy to spot would be x = 9/ with y = 0, that is (9/, 0). With no predators present, dx/dt is is just a logistic model without harvesting, initially at its carrying capacity and thus never leaving.

Things get interesting


What were really considered with here, is if our predators and prey are allowed to interact (nally some action). With a bit of back substitution (which can be veried by the reader), we nd a third equilibrium point 2 2, 3 3 . What this tells us is that it it is possible for our predator and prey to live in harmony. But the mere existence of such a point does nothing to tell us about the behavior of the system outside of this value, only that it exists.

Linearization
As it turns out, near its equilibrium point, this otherwise complicated system of equations will behave in ways similar to linear systems that we have previously studied. We can analyze this behavior analytically by considering the linearization of the system. We do this by computing the Jacobian matrix (9x x2 3xy) (9x x2 3xy) x 3x y = 9 2x + 3y y 2 + x (2y + xy) (2y + xy) x y

Linearization at equilibrium point 2, 3 2 3

If we then consider the linearization of the system at the equilibrium point 2, 3 2 , we end up 3 with 2 2 6 9 2x 3 3 3 3(2) = 2 2 3 3 2 + 2 3 3 0 Now that we have a linearization of our system near our equilibrium point, we can then closely approximate the behavior of system near the equilibrium point by computing the eigenvalues 0 = (2 )() + 18 4 = 2 + 2 + 18 4 then = 2 42 4(18 4) 2 2 + 4 18

Well this is interesting. Because weve chosen not to assign a value to we end up with eigenvalues that are dependent on alpha, which gives a quadratic with in a quadratic. Lets pause and take a look at what this tells us so far. 2

We rst notice that for = 0 our eigenvalues are going to be pure imaginary, and thus we can expect solutions to our system to be periodic with our equilibrium point acting as a center. We see a graphical representation of this in Figure 1.
y
5

3.5

=0 x(t) y(t)

2.5

1.5

0 0 1 2 3 4 5
0 1 2 3 4 5

(a) Periodic solutions

(b) x(t) and y(t), initial values: (3, 3)

Figure 1: = 0 As increases, but is close to 0, the terms under the radical will sum to a value less than 0, giving complex eigenvalues with a negative real part, which is indicative of a system with a spiral sink. Which we can see, graphically in Figure 2.
y
5

x t y t

=2 () ()

2.5

1.5

0 0 1 2 3 4 5

(a) A spiral sink

(b) x(t) and y(t), initial values: (3, 3)

Figure 2: = 2 As continues to increase, we will reach a point at which the sum of the terms under the radical are no longer negative. in this case will still be negative, meaning we will still have a sink, but close to our equilibrium point, it will behave less like a spiral, and approach equilibrium in a more linear fashion.

A quadratic within a quadratic With a little quadratic magic we can nd the roots of the polynomial under the radical in terms of . This will give the point at which our eigenvalues will cease to be complex. 16 4(18) 2 4 2 22 = 2 = 2 22 = 4 2 4.6904 Remember that we are only concerned with 0, we see that as increases, once it reaches a value of 2 + 22 2.6904 the terms under the radical will sum to a positive, and as mentioned, we will see x and y approach equilibrium more rapidly. We can observe this graphically in Figure 3.
y
5

x t y t

=3 () ()

2.5

1.5

0 0 1 2 3 4 5
0 1 2 3 4 5

(a) A not-so-spiral sink

(b) x(t) and y(t), initial values: (3, 3)

Figure 3: = 3

Bifurcation Values
The question that we have still neglected to answer is one of bifurcation. That is, a point that, regardless of initial conditions, our predator or prey will always become extinct. The computation thus far has been fairly involved, the answer to this, however happens to be quite easy to spot. If we consider the equilibrium point 2 2, 3 3 and draw our attention to the y value what we notice is that 9/2 that our y-value will be 0, which either means that predators dont exist, or does not make sense in the context of the problem. In other words, this is bad news bear. . . err, predators. As increases, the carrying capacity of our population model for our prey is decreases. Thus if one were a scientist concerned about the survival of a predator species3 one might start to be a little concerned as values start increasing since the predators food source is starting to run out. To see a graphical representation of this we rst consider a point close to bifurcation where = 4, as seen in Figure 4. Then we consider = 4.5, which is our point of bifurcation, as can be
3 or

if one were, like the author, concerned about a gardens kale production

seen in Figure 5. We note in both of these graphs, or x and y vs. t are considered over a very large interval of time. This is nessecery because we are so close to our point of bifurcation. Of course, for larger values of our predators would approach extinction much more rapidly.
y
2.8
2

2.4

2
1.5

1.6

x t y t
1.2
1

=4 () ()

0.8

0.4
0.5

0.4

0.8

1.2

1.6

2.4

2.8
0 5 10 15 20 25 30

(a) A low carrying capacity for the prey

(b) x(t) and y(t), initial values: (2, 2)

Figure 4: = 4

2.8
2

2.4

2
1.5

1.6

1.2
1

=4.5 x(t) y(t)

0.8

0.5

0.4

0.4

0.8

1.2

1.6

2.4

2.8
10 20 30 40 50

(a) Prey population cannot sustain predators

(b) x(t) and y(t), initial values: (2, 1.2)

Figure 5: = 4.5

Das könnte Ihnen auch gefallen