Anda di halaman 1dari 5

Numerical Methods for PDEs and Applications

Camile Fraga Delfino Kunz


18th October 2018

Biological models using Difference Equations


1 Cell Division
Suppose a population of cells divide synchronously, with each member producing
a daughter cells. We can write a simple equation to model the number of cells
in each generation:

Mn+1 = aMn

If we suppose that initially there are M0 cells, how big will be the population
after n generations?

Mn+1 = a(aMn−1 ) = a[a(aMn−2 )] = . . . = an M0

The growth of the population depends on the constant a,


• |a| > 1: Mn increases over successive generations;
• |a| < 1: Mn decreases over successive generations;
• a = 1: Mn is constant.

Linear difference equations of first order


The (linear) difference equation of first order relates the actual generation only
with the last one. The general solution of this kind of equation is given by
Mn = Cλn , where C > 0 is constant.

Mn+1 = aMn
Mn+1 − aMn = 0
Cλn+1 − aCλn = 0
Cλn (λ − a) = 0
λ=a

So we found that Mn = Can thought the general solution.


How we determine the constant C?

1
2 Propagation of annual plants
The annual plants leave their seeds on the end of summer. The flowering plants
will die, leaving their progeny in form of dormant seeds that must survive the
winter to give rise to the new generation. On the following spring a fraction
of this seeds germinates giving rise to the new generation. Some seeds might
remain dormant for one year or more before reviving. Others will just die due
to predation, disease or weather. In order for plants survive as a species, they
must generate a sufficiently large of population from year to year.
Consider that the seeds are produced at the end of their growth season on
August, and then they die. A fraction of these seeds survives the winter, and
some germinate on the beginning of the season on May, giving rise to new gen-
eration of plants. The germination fraction depends on the age of the seeds.
We consider that the seeds more than two-years-old are no longer viable.

Parameters
• γ: number of seeds per year on August;
• α: fraction of one-year-old seeds that germinates in May;
• β: fraction of two-years-old seeds that germinates in May;
• σ: fraction of seeds that survives a given winter.
We are going to at first attempt formulate the model using the following
variables:
• pn :number of plants in generation n;
• Sn1 :number of one-year-old seeds in April (before germination);
• Sn2 :number of two-years-old seeds in April (before germination);
• S̄n1 :number of one-year-old seeds left in May (after some have germinated);
• S̄n2 :num. of two-years-old seeds left in May (after some have germinated);
• Sn0 :number of new seeds produced in August.
Equations
In May, a fraction α of one-year-old and β of two-years-old seeds produce the
plants, thus:
pn = αSn1 + βSn2 (1)

The seed bank is reduced due to this germination, so the seeds left are:

S̄n1 = (1 − α)Sn1 (2)


S̄n2 = (1 − β)Sn2 (3)

In August, new seeds are produced at the rate of γ per year:

Sn0 = γpn (4)

2
Over the winter the seed bank changes by mortality and aging. Seeds that
are new in generation n will be old on the generation n + 1:
1
Sn+1 = σSn0 (5)
2
Sn+1 = σ S̄n1 (6)

Condensing the equations


The goal is to use the relations obtained to condense on a set of two equations
linking successive plants and seed generations. Substituting equation (4) in
equation (5) we get:
1
Sn+1 = σ(γpn ) (7)

Similarly for the equations (2) and (6) we get:


2
Sn+1 = σ(1 − α)Sn1 (8)

We rewrite equation (1) for generation n + 1:


1 2
pn+1 = αSn+1 + βSn+1 (9)

Now substituting in (9) the relations (7) and (8), and considering the equation
(7), we get the linear coupled system of two equations:

pn+1 = ασγpn + βσ(1 − α)Sn1
1 (10)
Sn+1 = σγpn

Note that we can simplify the system of two equations by substituting the seed
equation into the plants one, and we get a single linear difference equation of
second order :

pn+1 = ασγpn + βσ 2 (1 − α)γpn−1 (11)

Linear difference equations of second order


For difference equations of second order the actual generation depends on two
generations before:

xn + bxn−1 + cxn−2 = 0

Considering the general solution xn = Cλn , and substituting on the difference


equation, we get the following polynomial:

P (λ) = λ2 + bλ + c = 0

We have the following possible solutions depending on the roots of P,


• if λ1 6= λ2 : xn = C1 λn1 + C2 λn2
• if λ1 = λ2 = λ: xn = C1 λn + C2 nλn

3
• if λ1 , λ2 are complex: λ1 = a + bi and λ1 = a − bi and xn = C1 λn1 + C2 λn2 .
On the complex roots case, we can write:

xn = rn (A cos(θn) + B sin(θn))

with A = C1 + C2 , B = C1 − C2 and rn = a2 + b2 .

3 An insect population
Insects have generally more than one stage in their life cycle from progeny to
maturity. As an example, consider the poplar gall aphid. The adult female
aphids produce the galls. All progeny of a single aphid is contained on one gall
on the leaves on poplars. Some fraction of it will survive until adulthood. For
modeling their population, let’s consider the following parameters and variables:
• an : number of adults female aphids in the n-th generation;
• pn : number of progeny in the n-th generation;
• m: fractional mortality of the young aphids;
• f : number of progeny per female aphid;
• r: ratio of female to total adults aphids.
The model that describes number of insects (female) and progeny is given by:

pn+1 = f an
an+1 = r(1 − m)f an

Systems of linear difference equations


Consider the two equations system:

xn+1 = a11 xn + a12 yn
yn+1 = a21 xn + a22 yn

Writing the system as Sn+1 = ASn We can transform this system into a second
order difference equation, depending on only one variable:

0 = xn+2 − tr(A)xn+1 + det(A)xn

where tr(A) is the trace of the matrix A, det(A) its determinant. Now we can
we find it’s solution using the same approach studied before. We want to find
the roots of the polynomial function for λ which is given by:

P (λ) = λ2 − tr(A)λ + det(A) = 0

4
4 Exercises
It is encouraged to do this activity in groups of at maximum three people. It is
also fine if you want to do it alone.
1) Consider the Model 1 (cell populations).
a) Calculate the constant C considering the initial populations of cells M0 .
b) Make a Matlab routine to obtain the number of cells over the generations
using the difference equation. You can use N = 20 generations or any other
value. Test what happens for a = 1, |a| > 1 and |a| < 1. Does negative values
for a make sense biologically?
c) If we consider the steady-state (when the cells number doesn’t change along
generations), what value for a we would have?

2) Consider the Model 2 (propagation of annual plants).


a) Make an Matlab routine to calculate the plants populations and number of
seeds over the years for the system (10). Play with the parameters α, β, γ and
σ. Also play with the initial value of plants p0 or seeds S0 . Can you make the
plants survive?
b) Now consider the following set of parameters, and a fixed initial number of
plants or seeds for both:
i) α = 0.5, β = 0.25, γ = 2 and σ = 0.8;
ii) α = 0.6, β = 0.3, γ = 2 and σ = 0.8;
what do you observe on each case?
c) Use the equation of second order (11) to find the relation between λ and the
parameters.
d) For the the both sets of parameters on item b), calculate λ1,2 , what do you
observe?
e) Using the initial number of plants p0 on eq. (11), how to obtain the value of
the constants C1,2 at the general solution pn+1 = C1 λn1 + C2 λn2 ?
f ) If you are successful on the past item, plot the results on Matlab considering
the same set of parameters on b). How does it look like?

3) Consider the Model 3 (an insect population).


a) Discuss with your group if the model makes sense.
b) Using the theory for systems of linear difference equations, obtain the general
solution for the number of progeny at n − th generation.
c) Make a Matlab routine for the system of equations and play with the param-
eters.

References
1 Edelstein-Keshet, L. Mathematical Models in Biology, SIAM, 2005.

Anda mungkin juga menyukai