Logistic and Surge function investigation

Outline
The purpose of this investigation is to outline the properties and graphs of both surge and logistic functions, as well as investigating
their suitability for graphing certain scenarios. By listing several examples of the functions in real situations, the suitability of the
functions for real applications will be tested. The exact properties of the graph will be investigated by use of calculus.​

The Surge Function​

−𝑘𝑥
The surge function is given in form 𝑓(𝑥) = 𝐴𝑥𝑒 . This function models a graph showing a rapid increase followed by an exponential
decrease after the function reaches its peak. For this function, the subject to the exponential is always Euler’s number (𝑒). There are two
main variables that affect the shape of the surge function, 𝐴 and 𝑘, both affecting different parts of the function.
Surge function graphs

From testing several surge function graphs the general effects of the 𝐴 and 𝑘 variables can be concluded. The 𝐴 variable affects the height
of the maximum point of the function and by result affects the overall amplitude of the function.
As for the 𝑘 variable, it appears to have an inverse relationship with the height of the maximum point of the function and by result
affects the overall amplitude of the function. The 𝑘 variable also affects the maximum and inflection point on the graph along the 𝑥-axis,
as 𝑘 increases the maximum and inflection points to approach the 𝑦-axis.
Mathematical reasoning
Using calculus (specifically differentiation) on the surge function, the maximum and inflection points of the graph can be determined.
−𝑘𝑥
Using this differentiation on the general surge function (𝑓(𝑥) = 𝐴𝑥𝑒 ) the exact effect that the variables 𝐴 and 𝑘 have on the
maximum and inflection points of the graph, and by extension the overall shape of the graph, can be determined.
Refer calculations to appendix 1
The above calculations confirm the observations made from the graphing about the variables' effect on the function. Using the above
calculations, the specific effects can be identified.
It shows in the calculations that the 𝐴 variable has no effect on the 𝑥 coordinates of the maximum point of inflection point. This also
means that the 𝐴 variable has no effect on the horizontal dilation of the surge function. The 𝐴 variable does affect the vertical dilation of
the function, as shown by the calculations. The 𝐴 variable is directly proportional to the 𝑦 coordinates of the function, being the factor
the coordinate is multiplied by. This essentially makes the 𝐴 variables affect the height of the surge function.
It shows in the calculations that the 𝑘 variable is the factor the 𝑥 coordinates of the maximum point of inflection point is divided by.
This also means that the 𝑘 variable has an inverse proportional effect on the horizontal dilation of the surge function. The 𝑘 variable also
has an affect on the vertical dilation of the function, as shown by the calculations. The 𝑘 variable is inversely proportional to the 𝑦
coordinates of the function, being the factor the coordinates are divided by. This essentially makes the 𝑘 variables affect the height and
length of the surge function inversely.
Calculations, effect of variables in surge functions (Appendix 1)
−𝑘𝑥 −𝑘𝑥 −𝑘𝑥 −𝑘𝑥
𝑓(𝑥) = 𝐴𝑥𝑒 ⇒ 𝑓'(𝑥) = (𝑒 )(𝐴) + (𝐴𝑥)(− 𝑘𝑒 ) = 𝐴𝑒 (1 − 𝑘𝑥)
−𝑘𝑥 −𝑘𝑥 −𝑘𝑥 2 −𝑘𝑥 −𝑘𝑥
⇒ 𝑓''(𝑥) = (1 − 𝑘𝑥)(− 𝐴𝑘𝑒 ) + (𝐴𝑒 )(− 𝑘) =− 𝐴𝑘𝑒 + 𝐴𝑘 𝑥𝑒 − 𝐴𝑘𝑒
2 −𝑘𝑥 −𝑘𝑥 −𝑘𝑥
= 𝐴𝑘 𝑥𝑒 − 2𝐴𝑘𝑒 = 𝐴𝑘𝑒 (𝑘𝑥 − 2)
At maximum 𝑓'(𝑥) = 0
−𝑘𝑥 −𝑘𝑥
𝐴𝑒 (1 − 𝑘𝑥) = 0 ⇒ 𝐴𝑒 = 0, (1 − 𝑘𝑥) = 0
−𝑘𝑥 1
⇒ 𝐴𝑒 ≠ 0 ∴(1 − 𝑘𝑥) = 0 ⇒ 𝑥 = 𝑘
1 −1
−𝑘 𝑘 𝐴𝑒 1 𝐴
1
⇒ 𝑓( ) = 𝐴 𝑒
𝑘
1
𝑘
=
𝑘
∴ Max point = ( 𝑘 , 𝑘𝑒
)
At inflection point 𝑓''(𝑥) = 0
−𝑘𝑥 −𝑘𝑥
𝐴𝑘𝑒 (𝑘𝑥 − 2) = 0 ⇒ 𝐴𝑘𝑒 = 0, (𝑘𝑥 − 2) = 0
−𝑘𝑥 2
⇒ 𝐴𝑘𝑒 ≠ 0, (𝑘𝑥 − 2) = 0 ⇒ 𝑥 = 𝑘
2 −2
−𝑘 𝑘 2𝐴𝑒 2 2𝐴
2
⇒ 𝑓( ) = 𝐴 𝑒
𝑘
2
𝑘
=
𝑘
∴ Inflection point = ( 𝑘 , 2 )
𝑘𝑒

The Logistic Function

𝑎 𝐿
The logistic function is given in the basic form 𝑓(𝑥) = −𝑎𝑥 or the limit form 𝑓(𝑥) = −𝑘𝑥 . This function is used as a
𝑏+𝑐𝑒 1+𝐴𝑒

replacement for exponential growth functions in more real scenarios, such as population growth or disease spread. This is due to this
𝑏𝑥
function having an asymptote limit unlike the exponential function (𝑓(𝑥) = 𝑎 + 𝑐) allowing real scenarios to be accurately graphed.
The shape of this graph is that of an exponential increase followed by a gradual gradient decrease until the function asymptotes at a
𝑑𝑦 2
limit. This function is given by integrating the differential function of 𝑑𝑥
= 𝑎𝑦 − 𝑏𝑦 . Depending on how the calculation is finalised,
both the basic and limit form of the logistic function can be determined from this integration. For the following testing the limit form
of the logistic equation will be used due to it being easily understandable. It should be noted that the two forms are not separated by
𝑎
much, as the two forms can be easily interchanged with the fact that 𝐿 = 𝑏
.

Logistic function graphs

The graphs above show the properties that the variables of the logistic function affect. As the limit form is named, the 𝐿 variable affects
the limit point of the exponential, or the height at which the function can reach before asymptoting. The 𝐴 variable inversely affects the
height of the overall function, it doesn’t however affect the final asymptoting height of the function, meaning that a higher 𝐴 value may
cause the function to be overall lower it will still asymptote at a final height determined by 𝐿. The 𝑘 variable affects the gradient of the
function until the function begins to asymptote as determined by 𝐿.
Mathematical reasoning
Using calculus (specifically differentiation) on the logistic function, the inflection point of the graph can be determined. Using this
𝐿
differentiation on the limit form of the logistic function (𝑓(𝑥) = −𝑘𝑥 ) the exact effect that the variables 𝐿, 𝐴 and 𝑘 have on the
1+𝐴𝑒
inflection point of the graph, and by extension their effect on the overall shape of the graph, can be determined.
Refer to calculations appendix 2
The above calculation reveals the exact function of each variable in the limit form of the logistic function, both crediting and
discrediting some of the observed properties from the graphs.

The L variable has no effect on the 𝑥 coordinate of the inflection point, and therefore no effect on the horizontal dilation of the
function. The 𝐿 variable does affect the vertical dilation and asymptoting height of the function, as shown by the calculations. The 𝐿
variable is directly proportional to the 𝑦 coordinates of the function, being the factor the coordinate is multiplied by. This essentially
makes the 𝐿 variables affect the height of the logistic function.

The 𝑘 variable has no effect on the y coordinate of the inflection point, and therefore no effect on the vertical dilation of the function.
The 𝑘 variable is the factor the 𝑥 coordinates of the maximum point of inflection point is divided by. This also means that the 𝑘 variable
has an inverse proportional effect on the horizontal dilation of the logistic function.
This explains the 𝑘 variable’s effect on the gradient, as the inflection point is moved closer to the 𝑦-axis the gradient will stop increasing
and start decreasing at a lower 𝑥. Due to the function still needing to reach an asymptoting height at 𝐿, the gradient at 𝑥 = 0 needs to
start larger so that a lower 𝑥 inflection point does not prevent the function from reaching that asymptoting height. These facts
essentially make the 𝑘 variable affects the gradient of the logistic function directly.

The 𝐴 variable has an effect on the 𝑥 and 𝑦 coordinate of the inflection point, and therefore an effect on the horizontal and vertical
dilation of the function. The natural logarithm subject of 𝐴 is the factor the 𝑥 coordinates of the inflection point. This also means that
the 𝐴 variable has an direct proportional effect on the horizontal dilation of the logistic function. The 𝐴 variable is also part of the
denominator of the 𝑦 coordinate. While 𝑒 (Euler's number) to the power of natural logarithm subject of 𝐴 is quotient to a term in the 𝑦
coordinate denominator, this has a lesser effect on the overall inflection point than A’s effect on the denominator. This means that the 𝐴
variable has an inverse proportional effect on the vertical dilation of the logistic function, though it will still reach the asymptotic height
detailed by 𝐿.
Calculations, effect of variables in logistic functions (Appendix 2)
𝐿 −𝑘𝑥 −1
𝑓(𝑥) = −𝑘𝑥 = 𝐿(1 + 𝐴𝑒 )
1+𝐴𝑒
−𝑘𝑥
−𝑘𝑥 −𝑘𝑥 −2 𝐿𝐴𝑘𝑒
⇒ 𝑓'(𝑥) = 𝐿 ×− 1 ×− 𝐴𝑘𝑒 × (1 + 𝐴𝑒 ) = −𝑘𝑥 2
(1+𝐴𝑒 )
−𝑘𝑥 𝑘𝑥 2 𝑘𝑥 𝑘𝑥
𝐿𝐴𝑘𝑒 (𝑒 ) 𝐿𝐴𝑘𝑒 𝐿𝐴𝑘𝑒
⇒ 𝑓'(𝑥) = −𝑘𝑥 2 × 𝑘𝑥 2 = 𝑘𝑥 −𝑘𝑥 2 = 𝑘𝑥 2
(1+𝐴𝑒 ) (𝑒 ) (𝑒 (1+𝐴𝑒 )) (𝑒 +𝐴)
𝑘𝑥 2 2 𝑘𝑥 𝑘𝑥 𝑘𝑥 𝑘𝑥
(𝑒 +𝐴) (𝐿𝐴𝑘 𝑒 )−(𝐿𝐴𝑘𝑒 )(2𝑘𝑒 (𝑒 +𝐴))
⇒ 𝑓''(𝑥) = 𝑘𝑥 4
(𝑒 +𝐴)
𝑘𝑥 2 2 𝑘𝑥 2 2𝑘𝑥 𝑘𝑥 𝑘𝑥 2 𝑘𝑥 2 2𝑘𝑥
(𝑒 +𝐴) (𝐿𝐴𝑘 𝑒 )−2𝐿𝐴𝑘 𝑒 (𝑒 +𝐴) (𝑒 +𝐴)(𝐿𝐴𝑘 𝑒 )−2𝐿𝐴𝑘 𝑒
= 𝑘𝑥 4 = 𝑘𝑥 3
(𝑒 +𝐴) (𝑒 +𝐴)
2 2𝑘𝑥 2 2 𝑘𝑥 2 2𝑘𝑥 2 2 𝑘𝑥 2 2𝑘𝑥
𝐿𝐴𝑘 𝑒 +𝐿𝐴 𝑘 𝑒 −2𝐿𝐴𝑘 𝑒 𝐿𝐴 𝑘 𝑒 −𝐿𝐴𝑘 𝑒
= 𝑘𝑥 3 = 𝑘𝑥 3
(𝑒 +𝐴) (𝑒 +𝐴)
At inflection point 𝑓''(𝑥) = 0
2 2 𝑘𝑥 2 2𝑘𝑥
𝐿𝐴 𝑘 𝑒 −𝐿𝐴𝑘 𝑒 2 2 𝑘𝑥 2 2𝑘𝑥
⇒ 𝑓''(𝑥) = 0 ⇒ 𝑘𝑥 3 = 0 ⇒ 𝐿𝐴 𝑘 𝑒 − 𝐿𝐴𝑘 𝑒 =0
(𝑒 +𝐴)
2 2 𝑘𝑥 2 2𝑘𝑥 𝑘𝑥 𝑙𝑛(𝐴)
⇒ 𝐿𝐴 𝑘 𝑒 = 𝐿𝐴𝑘 𝑒 ⇒𝐴 =𝑒 ⇒𝑥 = 𝑘
𝑙𝑛(𝐴) 𝐿 𝐿
𝑓( 𝑘
)= −𝑘
𝑙𝑛(𝐴) = −𝑙𝑛(𝐴)
1+𝐴𝑒 𝑘 1+𝐴𝑒
𝑙𝑛(𝐴) 𝐿
∴ Inflection point at ( 𝑘
, −𝑙𝑛(𝐴) )
1+𝐴𝑒

Scenarios modelled by surge functions

Surge functions can be used to model a few real applicable scenarios. The main use for the surge function in real applications is that of
monitoring concentration of medicine in the bloodstream overtime after injection, though the surge function can also model the
acceleration of a vehicle as the vehicle reaches higher speeds. The following are two scenarios of how the surge function can model
medicine concentration.
Scenario 1
−0.25𝑡
If the amount of a particular drug in the bloodstream is modelled by 𝐶(𝑡) = 5𝑡𝑒 in mg where 𝑡 is in hours, what is the peak
amount of drug in the bloodstream and what time this occurs in hours? Furthermore, when does the amount of drug in the blood
appear to reach 0mg?

Model graph of scenario ↓

Max point at 𝐶'(𝑡) = 0
𝑑𝐶 −0.25𝑡 −.25𝑡 −0.25𝑡 𝑡
⇒ 𝑑𝑡
= 𝐶'(𝑡) = (𝑒 )(5) + (5𝑡)(−. 25𝑒 ) = 5𝑒 (1 − 4
)
−0.25𝑡 𝑡 −0.25𝑡 𝑡
⇒ 𝐶'(𝑡) = 0 = 5𝑒 (1 − 4
) ∴ 5𝑒 = 0, (1 − 4
) = 0 (Null factor law)
−0.25𝑡 𝑡 𝑡
⇒ 5𝑒 ≠ 0, (1 − 4
) = 0 ∴ (1 − 4
) = 0⇒𝑡 = 4
−0.25(4) −1
⇒ 𝐶(4) = 5(4)𝑒 = 20𝑒 ≈ 7. 36𝑚𝑔
∴ 4 hours after injection is maxim amount of drug in blood at 7.36mg
⇒ 𝐶(𝑡) ≈ 0 at 𝑡 ≈ 31 (Graphical observation) ∴ Drug amount reaches 0 mg 31 hours after injection.
Scenario 2
If a given drug is administered to a patient’s excretory system to treat a kidney related illness, it is recorded to have reached maximum
concentration of 20 ng/ml after 5 hours. If the drug is required to be readministered when it has no more effect at a concentration of 8
ng/ml, when does the drug need to be readministered, in hours and minutes to the nearest minute.
⇒Scenario involves medicine dosage ∴ modelled by surge function.
−𝑘𝑥 1 𝐴
⇒Recall that for general surge function (𝑓(𝑥) = 𝐴𝑥𝑒 ) max point at ( , )
𝑘 𝑘𝑒
1 𝐴 1 5𝐴 20𝑒
⇒ Max point at (5, 20) ∴ 𝑘
= 5, 𝑘𝑒
= 20 ⇒ 𝑘 = 5
⇒ 𝑒
= 20 ⇒ 𝐴 = 5
= 4𝑒 = 10. 87(𝑟.2𝑑𝑝)
1 −0.2𝑥
⇒𝑘 = 5
, 𝐴 = 4𝑒 = 10. 87(𝑟.2𝑑𝑝) ∴ 𝑓(𝑥) = 4𝑒𝑥𝑒
⇒ Graphical representation ↓

−0.2𝑥
⇒ 𝑓(𝑥) = 8 ⇒ 𝑆𝑜𝑙𝑣𝑒(4𝑒𝑥𝑒 = 8, 𝑥) ⇒ 𝑥 =. 88, 15. 11 (𝑟. 2𝑑𝑝)​​ (Classpad solve function)
⇒ First 𝑥 at approach to peak (graphical interpretation) ∴ 𝑥 = 15. 11 when concentration falls below 8 ng/ml
∴ 15.11 hours after initial injection drug needs readministering
⇒ 15. 11 − 15 = 0. 11 ℎ𝑜𝑢𝑟𝑠 = 60 × 0. 11 = 6. 69 𝑚𝑖𝑛𝑢𝑡𝑒𝑠
∴ 15 hours and 7 minutes after initial injection drug needs readministering
Scenarios modelled by logistic functions
Logistic functions can be used to model many real applicable scenarios. Some uses for the logistic function to model real applications
include: the spread of disease over time, population growth over time, spread of information in a group over time, concentration of a
substance in a chemical reaction & pH levels for a titration as volume of base is added. The following are two scenarios of how the
logistic function can model real scenarios.
Scenario 1
If a particular human population inhabiting a certain land can reach a maximum population of 2.76 million due to resource supply. It is
recorded that after 180 years the population is at 1.5 million people, and after 250 years the population is at 2.64 million. Using this
information, create a model for this population growth.
⇒Scenario is Population growth ∴ can be modelled by logistic function
⇒Function passes through points (1. 8, 1. 5) & (2. 5, 2. 64) (Letting 𝑡 be hundreds of years, letting 𝑦 be millions of people)
⇒Max population of 2.76 million ∴ 𝐿 = 2. 76
2.76
⇒ ​ −1.8𝑘 1. 5​ ⇒ 𝐴 = 1518. 843(𝑟.3𝑑𝑝), 𝑘 = 4. 167(𝑟.3𝑑𝑝)
1+𝐴𝑒
2.76
​ −2.5𝑘 = 2. 64 (Classpad simultaneous equations used)
1+𝐴𝑒
2.76
⇒ 𝑃(𝑡) = −4.167𝑡
1+1518.843𝑒
⇒ Graph of function ↓​ ​ ​ ​ ​ ​ Desmos input ↓
Scenario 2
If a particular titration of a known concentration of ethanoic acid as the primary standard aliquot is titrated against an unknown
concentration of sodium hydroxide, a particular titration curve will be created. If a buffer solution formed between ethanoic acid and
ethanoate resists any main change of pH until 22.5 ml of sodium hydroxide have been applied, what does the equivalence point of this
titration appear to be?
5𝑥−116
(Let the function of the buffer solution be 𝐵(𝑙) = 𝑥−23
{0 ≤ 𝑙 ≤ 22. 5} )
⇒Titration curve showing pH change can be modelled by logistic function.
⇒pH will maximise at pH 14 due to sodium hydroxide having a pH of 14, at ∞ ml of Sodium hydroxide pH will tend to 14 ∴ 𝐿 = 14
⇒Buffer equation will have same gradient and be at same point as as pH change equation at 𝑙 = 22. 5
(𝑥−23)(5)−(5𝑥−116)(1) 1
⇒ 𝐵'(𝑙) = 2 = 2
(𝑥−23) (𝑥−23)
1 5(22.5)−116
⇒ 𝐵'(22. 5) = 2 = 4, 𝐵(22. 5) = 22.5−23
=7
(22.5−23)
⇒Logistic equations tend to start at 𝑥 = 0 ≠ 22 ∴ let 𝑙 = 22 be 𝑥 = 0 ∴ 𝑥 = 𝑙 − 22
𝑘𝑥
14 14𝐴𝑘𝑒
⇒Let pH change equation be 𝑃(𝑥) = −𝑘𝑥 ⇒ 𝑃'(𝑥) = 𝑘𝑥 2
1+𝐴𝑒 (𝑒 +𝐴)
14
⇒​ −0.5𝑘 = 22. 5​ ⇒ 𝐴 = 1. 796(𝑟.3𝑑𝑝), 𝑘 = 1. 171(𝑟.3𝑑𝑝)
1+𝐴𝑒
0.5𝑘
14𝐴𝑘𝑒
​ 0.5𝑘 2 = 4​​ (Classpad simultaneous equations used)
(𝑒 +𝐴)
14
⇒𝑃(𝑥) = −1.171𝑥
1+1.796𝑒
14
⇒ 𝑥 = 𝑙 − 22 ∴ 𝑃(𝑥) = −1.171(𝑙−22)
1+1.796𝑒

⇒ pH of ethanoic acid 𝑙 = 0 ⇒ pH= 5. 04(𝑟.2𝑑𝑝)

⇒pH of sodium hydroxide pH= 14
14+5.04
⇒Equivalence point at average pH= 2
= 9. 52

14
⇒ 𝑃(𝑥) = −1.171(𝑙−22) = 9. 52 ⇒ 𝑙 = 23. 14(𝑟.2𝑑𝑝)
1+1.796𝑒
∴ Equivalence point at 23.14 ml
Conclusion
The findings of this investigation have revealed the overall properties of both the surge and logistic function as well as their uses in real
−𝑘𝑥
applicable scenarios. The surge function, given in the form 𝑓(𝑥) = 𝐴𝑥𝑒 , models a graph showing a rapid increase followed by an
exponential decrease after the function reaches its peak. There are two variables in the surge function, being 𝐴 and 𝑘. The 𝐴 variable
affects the height of the surge function, while the 𝑘 variable affects the height and length of the surge function inversely. The two main
real applications of the surge function are in the medicine field and in vehicle acceleration. In the medicine field, the surge function is
suitable to be used for monitoring of medicine/drug concentration or volume in the bloodstream over a period of time. In vehicle
acceleration, the surge function is applicable to be used to show the vehicle's specific acceleration at a given time.
𝐿 𝐿
The logistic function, given in form 𝑓(𝑥) = −𝑘𝑥 or 𝑓(𝑥) = −𝑘𝑥 , models a graph the shape of an exponential increase followed
1+𝐴𝑒 1+𝐴𝑒
by a gradual gradient decrease until the function asymptotes at a limit. This function is used as a replacement for exponential growth
functions in more real scenarios due to this function having an asymptote limit, allowing real exponential growth scenarios to be
accurately graphed. There are three variables in the limit form of the logistic function, being 𝐿, 𝐴 and 𝑘. The 𝐿 variable affects the height
of the logistic function, or the asymptoting height. The 𝐴 variable has an inverse proportional effect on the vertical dilation of the
logistic function and a direct proportional effect on the horizontal dilation of the logistic function. The 𝑘 variable affects the gradient of
the logistic function directly by shifting the inflection point closer 𝑥 = 0. There are several real applications of the logistical function.
This function can be a suitable model for real situations such as population growth over time, disease spread over time, information
spread over time, titration curves (pH at a given volume of added base).

Limitations​​ ​ ​ ​ ​ ​ Graphs of logistic functions with negated variables ↓

There are two main limitations that affect this investigation, that being the lack of
investigating negative variables and rounding error.
The more significant limitation is that of not investigating negative variables. This refers to
the fact that this investigation did not explore the effects of negating the 𝐴 and 𝑘 variables for
the surge function nor the 𝐿,𝐴 or 𝑘 variables for the logistic function in limit form. This is
due to different reasons for each function. For the surge function, the investigation states that
the 𝐴 and 𝑘 variables are positive constants. For the logistic function, negating any of the
variables prompts the function to display odd properties, such as displaying properties similar
𝑓(𝑥)
to single order 𝑔(𝑥)
functions or not mapping any points to the first quadrant of the
cartesian plane entirely. For these reasons, neither the surge or logistic functions have any real
application use when variables are negated, for that reason this topic has not been explored in
this investigation.​ ​ ​ ​ ​ ​ ​ ​ ​ Logistic scenario 2 rounding gap ↓
The second limitation is that of rounding error. In some of the shown scenarios, rounded values were
used for values that were too “unclean” in their exact form. For this reason, defined functions using these
rounded values may be slightly off their true value. One such example is shown on the right, where the
functions 𝐵(𝑙) & 𝑃(𝑙) from scenario 2 of the logistic equation scenarios, which were meant to intercept,
are actually 0.002 units apart. This error is negligible however, and does not cause massive issues in the
investigation.
Bibliography
Surge function and interactions, by Surge function introduction, by Desmos graphing calculator, last
Olta Tarko commondenominator viewed 14/8/24