You are on page 1of 9

MTH 256 - Lab 1

Justin Drawbert draw@pdx.edu February 6, 2012

Introduction
In this lab we consider logistic models of population growth that have been modeled in terms that account for harvesting. Our book purposes that we think of this in terms of a sh population in a body of water and shing. Given that the author does not eat sh and thus does not partake in the act of shing, this lab would pique the authors curiosity much more if we were to instead consider a human population p in some city. We then consider the harvest to be the occurrence of an alien species of walking sh-like creatures abducting humans. Presuming the humans in this city have no means of defense against said abductions, it would stand to reason that one might be quite interested in modeling the population with account to the rate of harvest.

1
1.1

Logistic growth with constant harvesting


First, the general case

We may nd it helpful to start by considering the population of a happy town without any sh monster abductions. We might model such a town with the logistic equation dp p = kp 1 dt N (1)

Perhaps the rst thing to notice, is that when p = 0, dp/dt will always be 0. This makes sense if you consider that if no humans were present initially, there will be no humans there to reproduce. Negative populations dont make much sense, given the context of the problem, so for the remainder of this paper we will assume p 0. We next address k and N which are known as the growth-rate parameter and the carrying capacity, respectively. The growth-rate will be just like it sounds, the rate at which our towns population tends to grow. The carrying capacity will then be the population at which our town will tend to, for whatever reason. Maybe old folks start to die after the population gets to large, or babies are born when the population gets to small. We note that for initial values of p > N that (1 p/N ) < 0 N and thus our population will decrease at a decreasing rate, approaching N as t . Similarly, for values of p such that 0 < p < N we see that (1 p/N ) > 0 and thus the population will increase (at a decreasing rate), approaching N as t . In the event that p = N , dp/dt = 0 and our population will stay a constant N . We also note that dp/dt will approach N regardless of the growth-rate parameter k, which will simply dictate how quickly it approaches N . In order to illustrate this, we make things easy and pick some pretty numbers, setting k and N both equal to 1. We may then graph the direction eld and plot a few solutions.

Here we see our equilibrium points at p = N = 1 and p = 0. We then see that for initial populations greater than 1, the population will decrease as time goes on, eventually approaching 1. Likewise for values greater than 0 and less than 1, the population will increase over time, approaching 1. Now to make things interesting, we consider the consider a situation in which sh monsters are abducting humans from our town and how it will eect the population over time. If the sh monsters are harvesting humans at a constant rate, we can then model the population, p, of the humans at a given unit time, t, with the autonomous equation p dp = kp 1 dt N a (2)

Here we note that since the equation is autonomous (that is, the right-hand side does not depend on t), the rate of harvest will be constant per unit time t. Rather than attempt to solve this equation for p by separating variables and then integrating (which is possible, but gets really complicated rather quickly), we will make use of the observations made about equation (1). Previously we saw points of equilibrium existed whenever dp/dt was equal to 0. Taking advantage of this we may investigate the behavior of the population with a constant rate of harvest a, by letting expressing the right-hand side of equation (2) as fa (p) = kp 1 We then let fa (p) = 0 and do a bit of algebra kp 1 p N a=0 p N a

kp2 + kN p aN = 0 Then, some quadratic magic gives the solution kN p= = N 2 (kN )2 4kaN 2k aN N2 4 k 2 (3)

We can now see that when

aN N2 < , 4 k a>

kN 4 fa (p) will have no real solutions and thus dp/dt will not have any equilibrium points. We then realize that for a = kN/4 we will only have one solution. For values of a < kN/4, we see that fa (p) will have 2 solutions. We can visualize this by xing k and N at 1, and adjusting a. We can then plot fa (p) (as is done in our textbook), visually conrming our observations. 1 0.5 0 fa (p) 0.5 1 1.5 2 2 1.5 1 0.5 0 p Plotting a few slope elds, we see how this will eect dp/dt. 0.5 1 1.5 2 2.5

which can also be expressed as

a > kN/4 a = kN/4 a < kN/4

(a) a < kN/4

(b) a = kN/4

(c) a > kN/4

For a < kN/4 we will have two points of equilibrium, the lesser being a source, and the greater a sink. Here, as long as the initial population is above equilibrium, the humans are safe and the population will tend toward the citys carrying capacity despite the sh alien abductions. However, should the initial population happen to be below this initial value when the abductions start to occur, we have what is known as a bifurcation and its bad news. . . humans. For a = kN/4, have one point of equilibrium that our population will either remain at if it begins there or decrease toward if is initially greater. For initial populations less than this, our city will eventually be vacant of humans. Lastly, we see the worst case, that is a > kN/4. This is the scenario in which youd either need to form a defense against the sh monsters (and thus attempting to lower the value of a) or consider moving to another city. 3

1.2

More specic values for k, N, and a.

So far the equations weve considered have all been very simple. We have set k and N equal to 1 and only varied the rate of harvest. With these general observations in mind, we will now consider what might happen if we were to vary the growth-rate k, N , as and a according to the following table (which we will reference for the remainder of the paper). Choice 1 2 k 0.25 0.50 N 4 2 a1 0.09 0.16 a2 0.25 0.25

Table 1: Values of k, N and a to consider

1.2.1

Choice 1

We start by letting a = a1 from choice 1 and setting our values according to the table above, giving p dp = 0.25p 1 dt 4 Right away, we can compute 0.09 (4)

kN 0.25 4 1 = = 4 4 4 Thus we see that a < kN/4. This is a good thing for our humans. Provided the initial population in the city is high enough, regardless of the abductions, we would expect our city to grow at a rate such that it will tend toward its carrying capacity. Its when the initial population is not sucient, that there is cause for. . . PANIC! We can nd these exact vales as we did before. We let the right-hand side be fa (p) = 0.25p 1 p 0.09, then solve for fa (p) = 0 by recalling (3). Doing so gives 4 p= 4 2 42 0.09 4 4 .25 36 4 25 8 5

=2 =2 p=

18 2 = 0.4 or p = = 3.6 5 5 Thus for initial populations lower than 2/5 units, we expect our city to not survive; the sh monsters will eventually have to nd another city to snack on. If our citys population starts at 2/5 units, it will remain there. Anything above this, and the citys population will grow or decrease towards its capacity of 8/5 population units. Plotting a direction eld should conrm this.

We can see by the direction eld, that our assumptions are indeed correct. 1.2.2 Choice 2

We next consider a city with a slightly smaller carrying capacity of 2 units, but with a faster growth rate. We also assume that our human harvesting sh aliens are harvesting humans at a greater rate of 0.16 per t. This gives the equation dp p = 0.50p 1 dt 2 0.16 (5)

As we have just done we make a quick check and nd that 0.5 2 1 kN = = 4 4 4 Here we again see that a < kN/4. Thus we expect our citys population to behave generally the same as before in that we will have two equilibrium points with one being a source and the other a sink. This time however, we notice that our citys growth rate is double and the carrying capacity half of what they were in Choice 1. The increase in growth rate, will presumably cause the citys population to approach equilibrium at a faster rate. And the decrease in carrying capacity, we expect will cause a decrease in these equilibrium values. We check this as we did before, by solving for our equilibrium values. p= 2 2 22 0.16 2 4 0.5 16 1 25 9 25

=1 =1 p=

8 2 = 0.4 or p = = 1.6 5 5 Plotting the direction eld, we see that our graph behaves as we hypothesized that it would. 5

Logistic growth with periodic harvesting

The previous model is a bit of a simplication of how things might go in an actual alien sh monster invasion. Suppose, that rather than at a harvesting at a constant rate there were other factors involved. Perhaps our sh monsters must return to their home planet periodically to refuel their ying shbowls as a result causing the harvest rate to vary sinusoidally. We could then model our population with p dp a (1 + sin bt) (6) = kp 1 dt N Here, p, t, k, and N are as they were before. p being the population at time t, k being the growthrate coecient, and N being the carrying capacity. a will also represent the rate of harvest, since a is being multiplied by (1 + sin bt). Since the sin function oscillates between -1 and 1, and thus the leading plus one ensures that our sh monsters are never contributing to the human population by adding humans to the city, causing the rate of harvest to be 0 when sin is at its minimum and 2a when sin is at its max. We also note that since the sine function completes its cycle every 2, letting b = 1 would mean that the cycle would complete every 2 units of t. If we were to say this in one sentence we could say that the rate of harvest will now vary sinusoidally between 0 and 2a with a frequency of b/(2) per unit t. Since our rate of harvest is now oscillating depending on time t, reaching a maximum of 2a, one might wonder the eect this will have on the values of equilibrium. Will it cause them to decrease? Or will the population be able to recover enough when harvest rate is decreasing that it wont matter? Since the equation is no longer autonomous, our previous methods of solving for points of equilibrium are not going to work. The equation is also no longer separable, so separation then integration isnt going to work either. Indeed solving for equilibrium values of p is going to be quite a complicated process and is beyond the scope of this paper. We can, however, investigate these equations graphically and see how they behave.

2.1

Choice 1

Referring to Table 1, we let a = a1 and b = 1, inputting parameters from Choice 1 into equation (6) gives p dp 0.09 (1 + sin t) (7) = 0.25p 1 dt 4 Noting that we have changed the scale of the graph to better t the equation, we look now to the slope eld for (7)

What is interesting here is that the equation appears to have the same equilibrium points as it did before, though in this case they are of course periodic.

2.2

Choice 2
p dp = 0.50p 1 dt 2

Here, we now consider Choice 2, letting a = a1 , giving 0.16 (1 + sin t) (8)

It is interesting to note that when we considered the constant rate of harvest with the same growthrate parameter and carrying capacity, we saw values of equilibrium at p = 0.4 and p = 1.6. However, here we note that because of the sine function there will be times when the rate of harvest is 2 0.16 = 0.32. For equal values of k and N with a constant rate of harvest of a = 0.32, we nd that a > kN/4 and thus there are no points of equilibrium (the humans are eventually doomed to extinction). This makes things interesting when considering what might happen here.

Interestingly we see the same trend here as we did in our previous example. We see that dp/dt oscillates about approximately same points of equilibrium as equation (5). This is interesting. Perhaps a way to begin to explain this might be considering that, for the constant harvest scenario, with k = 0.5 and N = 2 will only occur at values greater than kN/4 = 1/4. We then notice that the average value of a is .16 per 2 units of time. Thus the rate of harvest is only greater than 1/4 for values of t such that .16(1 + sin t) > 1/4, or whenever 1 + sin t > 1.5625. Continuing with that line of thought, it does not seem so unreasonable that dp/dt behaves the way it does. Though it would also stand to reason, that if you decreased the frequency enough, you could come up with a value for b such that there are no equilibrium points for the given parameters of Choice 2 with a = a1 (and indeed, this is the case. Say b = .1, for example).

Results of an increased rate of periodic harvesting

Here we consider the same functions as in the previous section, but with an increased rate of harvest, considering instead values of a2 from table 1.

3.1

Choice 1
dp p = 0.25p 1 dt 4

We rst consider the equation 0.25 (1 + sin t) (9)

Here we note that for for a constant growth rate of the same parameters a would be equal to kN/4. This would lead us to believe that we would have one equilibrium solution. It will be interesting to see what happens here

Here, we see that, at rst glance, things appear to be OK. However, if we take the same initial values and run t a little bit longer, we see that our population is, in the end, unsustainable.

3.2

Choice 2
p dp = 0.50p 1 dt 2

Lastly we consider Choice 2 with a = a2 giving 0.25 (1 + sin t) (10)

Here we note that just as in equation (9), a = kN/4. Will the greater growth rate save our city?

At rst glance, for the values weve chosen, things look good. But once again, taking t out a little further, and we see that our city is doomed even quicker than the previous example with a slower growth rate and a higher carrying capacity.

You might also like