Discord

INAIO Guide Stage 1 3Functions and Graphs

3.3 Exponential and log functions

Model repeated percentage change with an exponential and use a logarithm to recover its input.

A linear function adds the same amount at each step. An exponential function multiplies by the same factor at each step, so it can describe repeated growth or decay. A logarithm reverses that rule: it tells you how many multiplicative steps are needed to reach a value. This distinction appears in populations, decay, search, and quantities spanning very different scales.

This is intermediate material built on powers and roots, solving equations, and inequalities. Review OpenStax's exponential-functions section first, then its logarithmic-functions section. You should already be comfortable evaluating powers and reading a function graph.

A basic idea: multiply each time

Starting at 100100, multiplying by 22 each step gives 100,200,400,800,…100,200,400,800,\ldots. The rule is G(t)=100⋅2tG(t)=100\cdot2^t. Multiplying by 1/21/2 instead gives 100,50,25,12.5,…100,50,25,12.5,\ldots, or D(t)=100(1/2)tD(t)=100(1/2)^t. The first graph bends upward; the second approaches zero while staying positive.

Two side-by-side panels starting at 100 at step 0. The left growth curve doubles each step, reaching 800 at step 3. The right decay curve halves each step, reaching 12.5 at step 3.Two side-by-side panels starting at 100 at step 0. The left growth curve doubles each step, reaching 800 at step 3. The right decay curve halves each step, reaching 12.5 at step 3.
A constant factor produces curved growth or decay, even though the rule is simple.

A quick tour of formulas

From a repeated change to a rule

The common form is

f(t)=Abt,A>0,b>0,b≠1.f(t)=A b^t,\qquad A>0,\quad b>0,\quad b\ne1.

A=f(0)A=f(0) is the starting amount and bb is the multiplier per step. A growth rate of rr per step means b=1+rb=1+r: 5%5\% growth uses 1.051.05, so 100100 becomes 105105, then 110.25110.25. A decay rate of rr means b=1−rb=1-r for 0<r<10<r<1: 20%20\% decay uses 0.800.80, so 100100 becomes 8080, then 6464. Each percentage acts on the current amount. Adding or subtracting a fixed 55 each step would make a linear rule instead.

Rule What stays the same between equal steps Example from 100100
Linear: A+dtA+dt Difference dd 100,105,110100,105,110
Exponential: AbtAb^t Ratio bb 100,105,110.25100,105,110.25

Solve for the missing input

For b>0b>0, b≠1b\ne1, and y>0y>0, a logarithm is the exponent that makes bt=yb^t=y:

bt=y⟺t=log⁡by.b^t=y\quad\Longleftrightarrow\quad t=\log_b y.

When an amount is known, isolate the power first, just as Module 2.2 isolates an unknown in an equation:

Abt=T⟹bt=TA⟹t=log⁡b ⁣(TA)=ln⁡(T/A)ln⁡b.Ab^t=T \quad\Longrightarrow\quad b^t=\frac{T}{A} \quad\Longrightarrow\quad t=\log_b\!\left(\frac{T}{A}\right) =\frac{\ln(T/A)}{\ln b}.

The target TT must be positive. For a quick check, 50(1.2)t=86.450(1.2)^t=86.4 becomes (1.2)t=1.728(1.2)^t=1.728. Since 1.23=1.7281.2^3=1.728, t=3t=3. Taking a logarithm before dividing by 5050 would mix the multiplier with the power and obscure the equation.

Read the graph and the allowed inputs

An exponential with A>0A>0 accepts every real tt and gives positive outputs; its graph passes through (0,A)(0,A). If b>1b>1 it rises; if 0<b<10<b<1 it falls towards zero but never reaches zero. The logarithm reverses input and output, so it accepts only positive inputs. For b>1b>1, the points (1,b)(1,b) on y=bty=b^t and (b,1)(b,1) on y=log⁡bty=\log_b t are reflections across y=ty=t. The logarithm grows, but more slowly than the exponential.

When 0<b<10<b<1, both btb^t and log⁡bt\log_b t decrease. Therefore, applying log⁡b\log_b to both sides of an inequality reverses its direction, as in Module 2.4. If time is measured in complete rounds, use a whole-number answer and check the rounds just below and above the logarithmic boundary. A continuous answer such as t=3.2t=3.2 does not mean that three complete rounds are enough.

Application 1: medicine leaving a system

A simplified model says 8080 mg of a substance is present initially and half is left after each four-hour period. How much remains after 1212 hours?

  1. Count equal periods. Twelve hours contains 12/4=312/4=3 four-hour periods.
  2. Apply the same factor three times. M(12)=80(1/2)3=10M(12)=80(1/2)^3=10 mg.
  3. Check the sequence. The amounts after 44, 88, and 1212 hours are 4040, 2020, and 1010 mg. Subtracting 4040 mg every period would give a different, linear model.

This model describes the repeated fraction, not a fixed loss.

Application 2: yes-or-no questions

A question can have only two answers. In the best case, each question splits the remaining possibilities evenly. With the best choice of questions, what is the minimum number needed to distinguish among 10001000 possibilities?

  1. Count possible answer strings. With qq questions there are at most 2q2^q different sequences of answers.
  2. Set the requirement. We need 2q≥10002^q\ge1000, so q≥log⁡21000≈9.97q\ge\log_2 1000\approx9.97.
  3. Use a whole number and check. Nine questions distinguish at most 29=5122^9=512 cases; ten distinguish up to 210=10242^{10}=1024. The minimum is 1010.

The logarithm finds a continuous boundary. The question count is a whole number, so the final answer is the ceiling of that boundary.

Resources

SourceTitleWhy read it
OpenStaxExponential functionsSee growth, decay, and exponential models.
OpenStaxLogarithmic functionsLearn the inverse rule and read logarithmic graphs.

Finished this module? Mark it done.

Comments

Stuck on something, spotted a mistake, or found a resource that helped you? Tell us below.

Loading comments