Lesson 2-1: Advanced Exponential Functions & Modeling

Learning Objectives

Define and recognize exponential functions y = a^x and y = b e^{kx} with parameter meanings.
Determine growth vs. decay from base and rate parameters; interpret asymptotic behavior.
Compute compound interest and continuous compounding; compare compounding frequencies.

Use doubling time and half-life formulas to solve real-world problems.
Build and validate exponential models from context; interpret units and parameters.

Growth & Decay

Identify increasing/ decreasing behavior via base a and rate k; analyze end behavior and asymptotes.

Compounding

Discrete vs. continuous compounding; effective rates; financial applications.

Half-life

Decay modeling with physical meaning; relate to exponential constants and percent decrease.

Data Fit

Convert multiplicative change to additive logs for linearization and parameter estimation.

Time Scales

Doubling time T_d and half-life T_{1/2}; connect to k through natural logarithms.

Definition & Core Properties

An exponential function grows or decays by a constant percentage per equal time step. The two most common forms are base form and natural-base form:

y = a^x,\; a>0,\ a\neq 1

y = b\, e^{kx},\; b>0

Key features:

Domain: $(-\infty,\infty)$ ; Range: $(0,\infty)$ ; horizontal asymptote: $y=0$ .
If $a>1$ or $k>0$ , the function is increasing; if $0<a<1$ or $k<0$ , it is decreasing.
All exponential graphs pass through $(0,1)$ in base form; $(0,b)$ in natural-base form.

Growth vs. Decay & Percentage Interpretation

A convenient modeling form uses a growth factor per period:

y(t) = y_0 (1+r)^{t/\Delta t}

Here $r$ is the percentage change per base period $\Delta t$ . If $r>0$ , we have growth; if $r<0$ , we have decay.

Converting to natural-base form yields $y(t) = y_0 e^{kt}$ , where $k = \frac{\ln(1+r)}{\Delta t}$ . For small |r|, the approximation $k \approx \frac{r}{\Delta t}$ holds.

Compound Interest & Continuous Compounding

Financial growth often uses compounding. With nominal annual rate $r$ , principal $P$ , and $n$ compounding periods per year for $t$ years:

A = Pleft(1+ frac{r}{n} ight)^{nt}

As $n\to\infty$ , we obtain continuous compounding with base $e$ :

A = Pe^{rt}

Effective Annual Rate (EAR) compares different compounding frequencies on an equal footing:

EAR = left(1+ frac{r}{n} ight)^{n} - 1,quad EAR_{cont} = e^{r}-1

Doubling Time and Half-Life

For $y(t) = y_0 e^{kt}$ with $k>0$ (growth), the doubling time $T_d$ solves $e^{k T_d} = 2$ , hence:

T_d = frac{ln 2}{k}

For decay ( $k<0$ ), the half-life $T_{1/2}$ makes $e^{k T_{1/2}} = frac{1}{2}$ , so:

T_{1/2} = frac{ln 2}{|k|}

These relations allow quick reasoning about time scales in biology, physics, and finance.

Asymptotic Behavior & Transformations

Exponentials have horizontal asymptote $y=0$ . With vertical shift $+c$ , the asymptote becomes $y=c$ . Horizontal scaling modifies the time scale; vertical scaling changes initial value.

Vertical stretch: $y = A e^{kx}$ scales all values by $A$ .
Horizontal scaling: $y = e^{k(ax)}$ changes growth rate to $ka$ .
Vertical shift: $y = e^{kx} + c$ shifts asymptote to $y=c$ .

Worked Examples

Example 1: Compare Discrete vs. Continuous Compounding

Invest $P=10000$ at nominal annual rate $r=5\%$ for $t=5$ years. Compute with annual, monthly, and continuous compounding, and compare.

Annual (n=1): $A=10000(1+0.05)^{5}$ ≈ 12763

Monthly (n=12): $A=10000left(1+ frac{0.05}{12} ight)^{60}$ ≈ 12834

Continuous: $A=10000e^{0.05cdot 5}$ ≈ 12840

More frequent compounding yields higher returns; continuous compounding is the theoretical maximum for a given nominal rate.

Example 2: Radioactive Decay and Half-Life

A substance decays according to $m(t)=m_0 e^{kt}$ with half-life 30 days. Find $k$ and compute remaining mass after 20 days when $m_0=200$ g.

Half-life relation: $e^{kcdot 30}= frac{1}{2}$ ⇒ $k= frac{ln(1/2)}{30}=- frac{ln 2}{30}$

After 20 days: $m(20)=200left( frac{1}{2} ight)^{20/30}$ ≈ 126 g

The mass decreases by a constant percentage per equal time interval; the decay is multiplicative, not linear.

Example 3: Determine Doubling Time

A population follows $P(t)=P_0 e^{0.08t}$ with $t$ in years. Find the doubling time.

$T_d= frac{ln 2}{0.08}$ ≈ 8.66 years.

At 8% continuous growth, the quantity doubles about every 8.66 years.

Example 4: Effective Annual Rate (EAR)

Compare EAR for $r=12\%$ with monthly compounding vs. continuous compounding.

Monthly: $EAR=(1+0.12/12)^{12}-1$ ≈ 12.68%

Continuous: $EAR_{cont}=e^{0.12}-1$ ≈ 12.75%

Continuous compounding gives a slightly higher effective rate for the same nominal rate.

Modeling Scenarios

Finance: Savings Growth

With monthly deposits and monthly compounding, the future value follows an annuity-growth model. For simplicity, if the deposit occurs at period end, the closed form is:

FV = D, rac{left(1+ frac{r}{n} ight)^{nt}-1}{ frac{r}{n}}

This complements the single-sum formulas and is vital in budgeting and retirement planning.

Science: Pharmacokinetics

Drug concentration in blood often decays exponentially after reaching peak level. If the elimination constant is $k$ , the half-life is $T_{1/2}=ln 2/|k|$ . Repeated dosing superposes exponentials.

C(t)=C_0 e^{kt},quad k<0

Understanding half-life informs safe dosing intervals and steady-state analysis.

Earth Science: Cooling Laws

Newton's Law of Cooling often yields an exponential approach to ambient temperature. With ambient $T_a$ and initial $T_0$ :

T(t)=T_a + (T_0-T_a) e^{kt}

The asymptote shifts from 0 to $T_a$ , demonstrating transformation effects on exponential functions.

Demography: Population Growth

Short-run population growth can often be approximated by an exponential with rate $k$ . The doubling time provides an intuitive summary of pace.

P(t)=P_0 e^{kt},quad T_d=ln 2/k

In the long run, logistic models may be more appropriate, but exponential fits are useful locally.

Practice Problems

Problem 1: Determine Growth or Decay

Classify each function as growth or decay and state its asymptote:

$y=5cdot (0.7)^x$
$y=2e^{0.3x}$
$y=10 - 8cdot (0.5)^x$

Show Solution

(a) Decay (factor 0.7<1), asymptote y=0.

(b) Growth (k=0.3>0), asymptote y=0.

(c) Decay toward y=10 (vertical shift), asymptote y=10.

Problem 2: Continuous Compounding

Find $t$ such that $A=2P$ under continuous compounding with rate $r$ .

Show Solution

We solve $Pe^{rt}=2P$ ⇒ $e^{rt}=2$ ⇒ $t= frac{ln 2}{r}$ .

Problem 3: Find k from Two Data Points

A culture grows from 100 to 135 in 2 hours. Assume $N(t)=N_0e^{kt}$ . Find $k$ and the doubling time.

Show Solution

$frac{135}{100}=e^{2k}$ ⇒ $k= frac{1}{2}ln(1.35)$ ≈ 0.1508 h⁻¹.

$T_d=ln 2/k$ ≈ 4.596 hours.

Problem 4: Half-Life Application

A medication has half-life 6 hours. If the initial dose produces concentration 40 units, find concentration after 18 hours.

Show Solution

Three half-lives: $40cdot (1/2)^3=5$ units.

Problem 5: Determine Asymptote after Transformation

For $y=7-3e^{-0.4x}$ , state growth/decay and its horizontal asymptote.

Show Solution

Decay toward $y=7$ ; asymptote is $y=7$ as $x oinfty$ .

Key Takeaways

Exponential change is multiplicative with constant percentage per equal interval.
Continuous compounding uses $e^{rt}$ and typically yields the highest value for given nominal rate.
Doubling time and half-life connect directly to the exponential rate via natural logarithms.
Transformations shift asymptotes and scale initial values but preserve exponential character.

Log-Linearization Workshop

Turning Products into Sums

For data following $y=Ce^{kt}$ , take ln to get $ln y = ln C + kt$ (linear in t).

Parameter Estimation

Estimate slope k by linear regression on $(t,ln y)$ ; intercept gives $ln C$ .

Extended Applications

Finance: Mortgage Balance Decay

Model remaining balance with effective rate per period.
Compare discrete compounding vs. continuous approximation.
Discuss sensitivity to rate changes and payment cadence.

Epidemiology: Early-Phase Spread

Use $N(t)=N_0e^{kt}$ with growth constant linked to R₀.
Fit k from case counts; interpret doubling time.
Explain limits of exponential once interventions begin.

Ecology: Fish Population Growth

Estimate short-term growth rate from tagging surveys.
Compute time to double under favorable conditions.
Note logistic saturation beyond the short time horizon.

Physics: Capacitor Charging

Model voltage as $V(t)=V_0(1-e^{kt})$ with $k<0$ .
Identify time constant and 63% rise time.
Contrast with discharge modeled by pure exponential decay.

Chemistry: First-Order Reaction

Concentration $C(t)=C_0e^{kt}$ , estimate k via ln plot.
Relate half-life to $|k|$ and temperature effects.
Discuss deviations from first-order at high concentration.

Computer Science: Algorithmic Growth

Contrast exponential time $O(2^n)$ vs. polynomial.
Use doubling argument to estimate break-even sizes.
Explain why small input increases cause huge runtime jumps.

Economics: Inflation and Prices

Approximate CPI increases with continuous rate k.
Compute doubling time of prices at various k.
Discuss compounding effects on long-term planning.

Earth Science: Seismic Aftershock Decay

Short-run decay rates approximated exponentially.
Estimate half-life of aftershock intensity.
Explain transition to power-law for longer horizons.

Biology: Bacterial Culture

Estimate k from OD readings at equal intervals.
Compute generation time and compare media conditions.
Discuss limits due to nutrients and waste buildup.

Medicine: Tumor Volume Doubling

Use imaging data to estimate exponential growth.
Interpret doubling time with clinical caution.
Note transitions to Gompertz/logistic models later.

Energy: Battery Self-Discharge

Model loss with small negative k over months.
Estimate capacity after storage; compare chemistries.
Relate to temperature dependence of k.

Environmental: CO₂ Accumulation

Short-term exponential rise from emissions scenarios.
Compute time to reach given concentration threshold.
Discuss feedbacks that alter exponential trend.

Technology: User Growth

Fit k from weekly active users in early stage.
Translate to doubling time for planning capacity.
Warn about saturation and network effects later.

Practice Bank

Bank 1: Identify Form & Parameters

Given data, decide base vs. natural-base model.
Estimate $k$ or base $a$ with units.
State asymptote and intercepts.

Bank 2: Compounding Comparisons

Compute annual, monthly, and continuous values.
Report EAR and rank options.
Explain differences qualitatively.

Bank 3: Doubling/Half-Life

Find $T_d$ or $T_{1/2}$ .
Interpret in plain language.
Check feasibility bounds.

Bank 4: Fit from Two Points

Compute $k$ using two measurements.
Predict future value at time t.
Quantify percent error if data noisy.

Bank 5: Linearization

Plot $(t,ln y)$ and fit a line.
Extract slope/intercept to get parameters.
Check residuals for curvature.

Bank 6: Transformations

Apply vertical shifts and stretches.
Identify new asymptote and intercept.
Explain effect on parameters.

Bank 7: Rate Interpretation

Translate k into percent per unit time.
Contrast small vs. large k scenarios.
Provide a real-world analogy.

Bank 8: Mixed Compounding

Convert discrete to continuous rate.
Compare outcomes after fixed horizon.
Decide which is preferable and why.

Bank 9: Parameter Bounds

Given constraints, find allowed k range.
Test extremes and interpret.
State assumptions clearly.

Bank 10: Back-of-Envelope

Approximate changes using $e^{kDelta t}approx 1+kDelta t$ .
Check accuracy region.
Compare with exact computation.

Bank 11: Piecewise Contexts

Model two phases with different rates.
Ensure continuity at the switch time.
Interpret each phase separately.

Bank 12: Uncertainty in k

Provide interval for k from data CI.
Propagate to prediction interval.
Discuss decision implications.

Bank 13: Unit Consistency

Check units for k and time variable.
Convert units and recompute.
Explain changes to interpretation.

Bank 14: Parameter Solving

Solve for $y_0$ and k from two conditions.
State model uniquely.
Validate with third checkpoint.

Bank 15: Competing Models

Compare exponential vs. linear fit error.
Choose model by residual analysis.
Explain tradeoffs.

Bank 16: Forecasting

Provide 1-, 2-, 5-period forecasts.
Quantify uncertainty qualitatively.
Note limits of extrapolation.

FAQ (Extended)

Q: When should I prefer $e^{kt}$ over $a^x$ ?

Use $e^{kt}$ for naturally continuous change (continuous compounding, decay constants). Use $a^x$ for discrete-period change.

Q: How to estimate k robustly?

Prefer multi-point log–linear regression over two-point estimates; inspect residuals and outliers for model adequacy.

Back to Unit 2 Next Lesson 2-2