You are on page 1of 13

Investigation: Surge and Logistic Functions

SACE Number: 937651R

Intro/outline of problem

- Can surge and logistic functions be applied to real world scenarios?


- Investigating how functions are changed by their variables
- 4 scenarios to map out/compare
- What limitations are found
- Discussing suitability of surge and logistic functions for said problems

Part 1: Investigating Surge Functions


The general form of the surge function:
−bx
f ( x )= Ax e
Where A and b are positive constants

Finding f’(x) Finding f’’(x)


' −bx −bx '' −bx −bx −bx
f ( x )= A (e −bx e ) f ( x )=− A be −Ab (e −bx e )
' −bx '' −bx −bx 2 −bx
f ( x )= A e (1−bx ) f ( x )=− Ab e − Ab e + A b x e
'' −bx
f ( x )=Ab e (bx−2)

Investigating the effect of A

Graph 1: f(x), f’(x)’ and f’’(x) graphed when A = 10 and b = 4.

Marked on the graph


are the maximum of f(x)
(0.25, 0.92), and the x
intercepts of f’(x) (0.25)
and f’’(x) (0.5) showing
the x coordinates of the
turning point and the
point of inflection
respectively.

Graph 2: f(x), f’(x), and f’’(x) graphed when A = 5 and b = 4

Marked on the graph are


the maximum of f(x) (0.25,
0.46), and the x intercepts
of f’(x) (0.25) and f’’(x) (0.5)
showing the x coordinates
of the turning point and
the point of inflection
respectively.
Graph 3: f(x), f’(x), and f’’(x) graphed when A = 15 and b = 4

Marked on the graph are


the maximum of f(x) (0.25,
1.38), and the x intercepts
of f’(x) (0.25) and f’’(x) (0.5)
showing the x coordinates
of the turning point and the
point of inflection
respectively.

Graph 4: f(x), f’(x), and f’’(x) graphed when A = 20 and b = 4

Marked on the graph are


the maximum of f(x) (0.25,
1.839), and the x intercepts
of f’(x) (0.25) and f’’(x) (0.5)
showing the x coordinates of
the turning point and the
point of inflection
respectively.

Table 1: Coordinates of the stationary point and point of inflection for f(x) for different values of A

Value of A X coordinate of Y coordinate of X coordinate of point Y coordinate of point


stationary point stationary point of inflection of inflection
5 0.25 0.46 0.5 0.34
10 0.25 0.92 0.5 0.68
15 0.25 1.38 0.5 1.02
20 0.25 1.84 0.5 1.35
It can be observed as A increases by 5, f(0.25) increases by 0.46, and f(0,5) increases by 0.34, therefore as A
increases by 1, f(0.25) increases by 0.092, and f(0.5) increases by 0.67. the x values of these coordinates do not
change, therefore A only affects the range of f(x)
Investigating the effect of B

Graph 5: f(x), f’(x)’ and f’’(x) graphed when A = 10 and b = 4.

Marked on the graph are the


maximum of f(x) (0.25, 0.92),
and the x intercepts of f’(x)
(0.25) and f’’(x) (0.5) showing
the x coordinates of the
turning point and the point
of inflection respectively.

Graph 6: f(x), f’(x)’ and f’’(x) graphed when A = 10 and b = 2.

Marked on the graph are the


maximum of f(x) (0.5, 1.839),
and the x intercepts of f’(x)
(0.5) and f’’(x) (1) showing the
x coordinates of the turning
point and the point of
inflection respectively.

Graph 7: f(x), f’(x)’ and f’’(x) graphed when A = 10 and b = 3.

Marked on the graph are


the maximum of f(x) (0.333,
1.226), and the x intercepts
of f’(x) (0.333) and f’’(x)
(0.667) showing the x
coordinates of the turning
point and the point of
inflection respectively.
Graph 8: f(x), f’(x)’ and f’’(x) graphed when A = 10 and b = 5.

Marked on the graph are the


maximum of f(x) (0.2,
0.736), and the x intercepts
of f’(x) (0.2) and f’’(x) (0.4)
showing the x coordinates of
the turning point and the
point of inflection
respectively.

Table 2: Coordinates of the stationary point and point of inflection for f(x) for different values of b

Value of b X coordinate of Y coordinate of X coordinate of point Y coordinate of point


stationary point stationary point of inflection of inflection
2 0.5 1.839 1 1.353
3 0.333 1.226 0.667 0.902
4 0.25 0.92 0.5 0.68
5 0.2 0.736 0.4 0.541
1 2
The table above shows the x coordinate of the stationary point is , and the x value of the point of inflection being ,
b b
as b increases, the x values of the stationary point and point of inflection decrease.

Forming and Testing conjectures


−bx
Based on the data in table 2 of f ( x )= Ax e , it can be assumed that for all b, the x coordinate of the stationary
1
point is . Visually, this can be observed of the graph of f(x) for any value of b, even as b becomes negative
b
(providing b≠0).

This can be proved algebraically using the derivative of f(x), as the stationary point of f(x) is a zero of f’(x). Previously
it was shown that f ' ( x )= A e−bx (1−bx ), using this x can be found in terms of b by rearranging the equation.

Finding the x value of the stationary point of f(x)


Let A=10 and f’(x)=0
−bx
0=10 e (1−bx)
−bx
0=e (1−bx )
0
−bx
=1−bx
e
0=1−bx
bx=1
1
x=
b

Similarly, other previous observations can be proved by using f’(x) and f’’(x).
Based on the data in table 2, since the x value of the point of inflection of f(x) corresponds to the x intercept of f’’(x)
the value can be found algebraically in terms of b. Previously f’’(x) was found to be f '' ( x )=Ab e−bx (bx−2).

Finding the x value of the stationary point of f(x)


Let A=10 and f’’(x)=0
−bx
0=10 b e ( bx−2)
0
−bx
=(bx−2)
Ab e
0=bx−2
0+2=bx
2=bx
2
=x
b

In order to find the y values of the stationary point and point of inflection a similar method can be used. Letting b=4
1 2
and x= or x= respectively, the value of y in terms of A can be found.
b b
Finding the y value of the stationary point of f(x) Finding the y value of the point of inflection of f(x)
1 2
Let x= Let x=
b b
−bx −bx
y= Ax e y= Ax e
1 2
1 −b b 2 −b b
y= A e y= A e
b b
A −1 2 A −2
y= e y= e
b b

Table 3: comparing y value in terms of A, to the results of Table 1:

Value of A Y coordinate of A −1 Y coordinate of point 2 A −2


stationary point y= e of inflection y= e
4 4
5 0.45985 0.45985 0.33834 0.33834
10 0.91970 0.91970 0.67668 0.67668
15 1.37955 1.37955 1.01501 1.01501
20 1.83940 1.83940 1.35335 1.35335

As shown by comparing the data, the y coordinates of the stationary point and the point of inflection are given by
A −1 2 A −2
y= e and y= e .
4 4
Part 2: Investigating Logistic Functions
The general form of the logistic function

L
P (t)= −bx
1+ A e
Where A, b, and L are constants and x ≥ 0.

Finding P’(x) Finding P’’(x)


L
P ( x )= −bx
1+ A e
−bx ' A Lbe
P ( x )= 2
( 1+ Ae−bx )
0 ( 1 A e−bx )−L(−bA e−bx ) e
−bx
P ' ( x )= '
P ( x )= ALb ×
−bx 2
( 1+ A e ) ( 1+ Ae−bx )
2

2
'' ( −b e−bx ) ( 1+ A e−bx ) −2 e−bx (1+ A e−bx )(− Ab e
' A Lbe
−bx P ( x )= ALb× ¿¿¿
P ( x )=
−bx 2
( 1+ Ae )
2
ALb(−b e
−bx
( Ab e−bx +1 ) −2 ( A e−bx +1 ) (−Ab e−bx ) e
P ' ' ( x )= 4
( 1+ A e−bx )

Investigating the effect of L Graph 9: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 10, and b = 4.

Marked on the graph are the maximum of f’(x) (0.576, 1), and the x intercept of f’’(x) (0.576)
showing the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 1 meaning there is no ‘maximum’ of the equation, this
is represented by f’(x) having a horizontal asymptote of y = 0.

Graph 10: f(x), f’(x)’ and f’’(x) graphed when L = 2, A = 10, and b = 4.

Marked on the graph are the maximum of f’(x) (0.576, 2), and the x intercept of f’’(x) (0.576)
showing the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 2 meaning there is no ‘maximum’ of the equation, this
is represented by f’(x) having a horizontal asymptote of y = 0.

Graph 11: f(x), f’(x)’ and f’’(x) graphed when L = 3, A = 10, and b = 4.

Marked on the graph are the maximum of f’(x) (0.576, 3), and the x intercept of f’’(x) (0.576) showing
the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 3 meaning there is no ‘maximum’ of the equation, this is
represented by f’(x) having a horizontal asymptote of y = 0.

Graph 12: f(x), f’(x)’ and f’’(x) graphed when L = 4, A = 10, and b = 4.

Marked on the graph are the maximum of f’(x) (0.576, 4), and the x intercept of f’’(x) (0.576)
showing the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 4 meaning there is no ‘maximum’ of the equation, this
is represented by f’(x) having a horizontal asymptote of y = 0.

Table 4: the horizontal asymptote and coordinates for the point of inflection for different values of L

Value of L

1
2
3
4
As shown by table 4, the x coordinate of the point of inflection remains unchanged as L takes different
values. The horizontal asymptote of the graph has a y value equal to the value of L, and the y coordinate
of the point of inflection (f(0.576)) increases by 0.500 as L increases by 1. Therefore, the value of L affects
the range of the graph only.

Investigating the effect of b

Graph 13: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 10, and b = 2.

Marked on the graph are the maximum of f’(x) (1.151, 0.5), and the x intercept of f’’(x) (1.151) showing the
x coordinate of the point of inflection.
There is also a horizontal asymptote at y = 1 meaning there is no ‘maximum’ of the equation, this is represented by
f’(x) having a horizontal asymptote of y = 0.

Graph 14: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 10, and b = 4

Marked on the graph are the maximum of f’(x) (0.576, 1), and the x intercept of f’’(x) (0.576) showing
the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 1 meaning there is no ‘maximum’ of the equation, this is
represented by f’(x) having a horizontal asymptote of y = 0.

Graph 15: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 10, and b = 6

Marked on the graph are the maximum of f’(x) (0.384, 1.5), and the x intercept of f’’(x) (0.384)
showing the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 1 meaning there is no ‘maximum’ of the equation, this is
represented by f’(x) having a horizontal asymptote of y = 0.

Graph 16: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 10, and b = 7

Marked on the graph are the maximum of f’(x) (0.329, 1.75), and the x intercept of f’’(x) (0.329) showing
the x coordinate of the point of inflection.

There is also a horizontal asymptote at y = 1 meaning there is no ‘maximum’ of the equation, this is
represented by f’(x) having a horizontal asymptote of y = 0.

Table 5: the horizontal asymptote and coordinates for the point of inflection for different values of b

Value of b

2
4
6
7
Table 5 shows that as the value of b increases, the x coordinate of the point of inflection decreases by a decreasing
amount, while it cannot be determined visually, the x coordinate of the point of inflection can be determined
algebraically. Notably, the y coordinate of the point of inflection remains constant at 0.5, therefore the value of b
does not affect the domain or range of f(x).

Investigating the effect of A Graph 17: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 5, and b = 4

Marked on the graph are the


maximum of f’(x) (0.402, 1),
and the x intercept of f’’(x)
(0.402) showing the x
coordinate of the point of
inflection.

There is also a horizontal


asymptote at y = 1 meaning
there is no ‘maximum’ of the
equation, this is represented
by f’(x) having a horizontal
asymptote of y = 0.

Graph 18: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 10, and b = 4

Marked on the graph are the


maximum of f’(x) (0.576, 1),
and the x intercept of f’’(x)
(0.576) showing the x
coordinate of the point of
inflection.

There is also a horizontal


asymptote at y = 1 meaning
there is no ‘maximum’ of the
equation, this is represented
by f’(x) having a horizontal
asymptote of y = 0.

Graph 19: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 15, and b = 4

Marked on the graph are the


maximum of f’(x) (0.677, 1),
and the x intercept of f’’(x)
(0.677) showing the x
coordinate of the point of
inflection.

There is also a horizontal


asymptote at y = 1 meaning
there is no ‘maximum’ of the
equation, this is represented
by f’(x) having a horizontal
asymptote of y = 0.
Graph 20: f(x), f’(x)’ and f’’(x) graphed when L = 1, A = 16, and b = 4

Marked on the graph are the


maximum of f’(x) (0.693, 1),
and the x intercept of f’’(x)
(0.693) showing the x
coordinate of the point of
inflection.

There is also a horizontal


asymptote at y = 1 meaning
there is no ‘maximum’ of the
equation, this is represented
by f’(x) having a horizontal
asymptote of y = 0.

Table 6: the horizontal asymptote and coordinates for the point of inflection for different values of A

Value of A Horizontal asymptote


X coordinate of point of Y coordinate of point of
inflection inflection
5 1 0.402 0.500
10 1 0.576 0.500
15 1 0.677 0.500
16 1 0.693 0.500
As the value of A increases, the x coordinate of the point of inflection increases by a decreasing amount. Similar to
table 5, the y coordinate of the point of inflection remains constant at 0.5. The domain and range of f(x) remain
unaffected.

Forming and Testing Conjectures

While a pattern cannot be found through visually comparing values, it can be inferred that the x value of the point of
inflection is of the form ln(n) with n being some variable within f(x) due to the presence of exponential e.

Similar to the process of testing conjectures in part A, the coordinates of the point of inflection can be proved
algebraically in terms of both A and b, as the x intercept of f’’(x) corresponds to the point of inflection of f(x)

Finding the x value of the point of inflection in terms of A and b 2 2 −2 bx 2 −2 bx 2 2 −2 bx


A b e + A b e =2 A b e
Let L=1, and f’’(x)=0 2 −bx 2 2 −2 bx
2 2 −2bx 2 −bx Ab e =A b e
2 A (1)b e A ( 1) b e
0= 3
− 2
−bx
e =Ae
−2 bx

( Ae−bx +1 ) ( Ae−bx +1 ) −2bx


2 2 −2 bx
ln ⁡(e ¿¿−bx)=ln ⁡( A e )¿
0 2A b e 2 −bx −2 bx
= − Ab e ln ⁡(e ¿¿−bx)=ln ( A ) + ln ⁡(e )¿
( Ae + 1 ) ( Ae +1 )
−bx 2 −bx
−bx=ln ( A )−2 bx
2 2 −2 bx
2 −bx 2A b e bx=ln A
Ab e =
( Ae−bx +1 ) ln A
x=
Ab2 e−bx ( Ae−bx +1 ) =2 A 2 b2 e−2bx b
Since the value of L is shown to have no effect of the x value of the point of inflection it is appropriate to give L a
value when determining the value of x. Regardless of what value is given for L, including none, L will cancel out
leaving x in terms of A and b only.

Since the value of L does affect the range of the graph of f(x) it subsequently will affect the y value of the point of
inflection. This means that in order to find the y value of the point of inflection, L must be included as a variable.

L
Using the original function y= −bx the y value of the point of inflection can be found.
1+ Ae
Finding the y value of the point of inflection in terms of A
and b
ln A
Let x=
b
L
y= −bx
1+ Ae
L
y= lnA
−b
1+ Ae b
L
y= −lnA
1+ Ae
L
y=
1+1

L
y=
2

Part 3: Modelling and Applications

Modeling Medicinal doses using surge functions

Within the last 20 years, mathematicians have been researching the suitability of surge functions for mapping how
the concentration of medicinal drugs in the body change over time. From a glance the drug response curve appears
to resemble a surge function, with a rapid increase to a maximum, then a slow in comparison decrease.

Graph 21: The drug response curve of a 100mg dose of Viagra. Graph 22: An example of a surge function
Source: Pfizer Labs, 2010

It can be seen by comparison that both graphs are similar in shape, although some noticeable differences can be
observed. In theory, the purpose of modeling medicine dosages with surge functions is to monitor when the
effectiveness of a drug decreases and estimate when to provide the next dose to the patient. The first step to
observing the suitability, was to attempt to map the data to a surge function algebraically. Ideally a perfect match
would have been found, showing a direct correlation between the concentration of a drug, and a surge function.
Viagra was chosen as the drug of choice due to data being relatively accessible online. The points noted on (graph
21) were estimated, as a true value was not given. Values shown in Table 7.

Table 7: Estimated Data Values on Concentration of Viagra given by Graph 21

t(hours) 0.13 0.38 0.51 1.11 1.6 2.1 3.12 4.06 6.02 8.05 10.05 12.04 18.03 24
C(ng/mL) 1 50 325 445 415 358 255 180 82 56 39 25 15 10

A key point to note is the maximum concentration given (Cmax) and the time it occurs (tmax). This gives the local
maximum of the equation at the point (1.11, 445). Using this data, a surge function was fitted. Given that the
1 A −1
maximum of a surge function was found to be at the point ( , e ), A and b can be found to complete the function
b b
Axe-bx.

Finding the Value of b Finding the Value of A


1 100
1.11= Let b =
b 111
1.11b=1 A −1
445= e
1 b
b=
1.11 100 −1
445 × =A e
100 111
b=
111 44500
111
−1
=A
e
44500
e=A
111
100
44500 1− 111 x which were then graphed to in theory give the equation of the
These values give the equation f ( x )= xe
111
drug response curve of (Graph 21)

Viagra Concentration (ng/mL) Graph 23: Graph of proposed surge function of


drug response curve compared with actual results.

Both the proposed surge function and the actual


data points were graphed on the same axes to
provide a direct comparison between the two.
(Graph 22) shows that the two functions are not
identical, this could be due to limited data. During
the ‘peak’ of both functions data values appeared
to line up closely between t=0.51 and t=2.1
however after t=2.1 the functions obviously
differed with the proposed function decreasing at
a much faster rate than the drug did.

Should the second dose of the drug be


administered solely based on the proposed
function, an overdose would occur.
Time (hours)
Some limitations seen while using this model include the clearly not accurate function, this is likely to be the case as
the ‘maximum point’ used is likely to not be the true maximum
concentration obtained, without constant monitoring of blood concentration a more accurate maximum is unlikely to
be observed. This in-turn effects the values of A and b as this graph was solved under the assumption that t=1.11
was the true maximum concentration.

Another limitation to using a surge function to map the drug response curve of medicine is that it vastly
oversimplifies the complexity of pharmacokinetics, as well as not considering the differences in individuals’ genetics,
metabolism, other medications being taken, and medical conditions that may affect drug absorption. Further, a
surge function has a horizonal asymptote of y=0 meaning the function will never reach zero unlike the human body,
where eventually all the drug would be excreted. If surge functions were to be standardized in graphing drug
concentrations, then a ‘cut off point’ would have to be established, at what point does a number get so small that it
is equal to zero?

Finally, this model only accounts for single dose treatment, where a second dose is only administered once the first
has worn off. Not all drugs are administered in this way, certain pain medications are administered by continuous
intravenous infusion, a continuous drip rather than one initial dose, which would result in a completely different
looking drug response curve that might not even be comparable to a surge function.
In conclusion a surge function may prove useful in specific situations with more research, however it is not a ‘one
size fits all’ and needs to be catered to the individual patient, their biological factors, and should be used in
conjunction with other modeling methods. The limitations shown are far too great to be the main source of
administration timing.

Limited Population Growth Models using logistic functions.

Graph 24: World population recorded since 1950 and


predicted population until 2100

Source: World Population Prospects 2022, United Nations


Population Division

The global population has rapidly increased as a result of


the industrial revolution, a sum that has continued to
increase throughout the 20 th century, reaching 8 billion in
2022. The United Nations regularly make predictions
about how the population will grow, usually with
incredible accuracy in short to mid-term projections
(Ritchie, 2023).

Previously it was suggested that the population would


continue to increase well past 12.3 billion (United
Nations, 2012), with more recent predictions suggesting a
maximum occurring before the year 2100. The most recent prediction (2022) from United Nations suggests the
population will reach a maximum of 10.4 billion in the year 2086, while other studies predict a smaller maximum
such as The Lancet estimating the population will peak at 9.73 billion in 2064 (Vollset, S. Et al. 2020).

Visually, the graph of the predicted population shares similar qualities to the graph of a logistic function, with a
slow rise pre-industrial revolution, a rapid increase as technology and medicine advances, and once again slowing
as the population reaches its maximum.

Similar to modelling medicinal dosing, a surge function was estimated based on the given information in the graph
provided by United Nations. L was taken as the maximum population given (10.4b). Taking some known points
such as the population milestones in graph 24, A and b can be found.

- Factors that affect population


o Migration
o Mortality
o Longevity
o Available resources (food, water, space, etc.)
o Fertility
o Women’s rights
o Access to birth control
o Cultural norms
- Limitations
o Population is expected to fall rather than remain consistent.
o Predictions aren’t accurate for long term
o Assumes an easily predictable birth and death rate

References

Gordon, S. 2006. FITTING SURGE FUNCTIONS TO DATA. Problems, Resources, and Issues in Mathematics
Undergraduate Studies.
https://www.tandfonline.com/doi/pdf/10.1080/10511970608984137
Peek, K. 2022. Global Population Growth Is Slowing Down. Here’s One Reason Why. Scientific American.
https://www.scientificamerican.com/article/global-population-growth-is-slowing-down-heres-one-reason-why/

Pfizer Labs. 2010. Viagra. Food and Drugs Administration.


https://www.accessdata.fda.gov/drugsatfda_docs/label/2010/020895s033lbl.pdf

Ritchie, H. 2023. The UN has made population projections for more than 50 years – how accurate have they been? Our
World in Data.
https://ourworldindata.org/population-projections

Tarko, O. 2021. Surge Functions and Drug Interactions. Undergraduate Journal of Mathematical Modelling: One +
Two. https://digitalcommons.usf.edu/cgi/viewcontent.cgi?
article=4942&context=ujmm

United Nations, 2022. World population to reach 8 billion on 15 November 2022. United Nations, Department of
Economic and Social Affairs.
https://www.un.org/en/desa/world-population-reach-8-billion-15-november-2022

Vollset, S. Et al. 2020. Fertility, mortality, migration, and population scenarios for 195 countries and territories from
2017 to 2100: a forecasting analysis for the Global Burden of Disease study. The Lancet.
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30677-2/fulltext

You might also like