Right end: \(x\to+\infty,\ y\to+\infty\)
Chapter 5 · Exponential Functions
指数函数 — Properties, Transformations, Modeling & Validation
- 掌握指数函数的图像与性质 · Know graphs and properties of exponential functions
- 判断并计算指数函数的端点行为 · Evaluate end behavior using limit notation
- 运用指数法则化简表达式 · Apply laws of exponents
- 理解水平平移与垂直伸缩的等价关系 · Connect horizontal shifts to vertical dilations
- 建立指数增长/衰减模型解决实际问题 · Build exponential growth/decay models
- 由两点、平移($ab^x+k$)和指数回归建立模型;换时间单位改写 · Build models from two points, shifts, and ExpReg; rewrite for other time units
- 用残差图验证模型,并判断模型是高估还是低估 · Validate models with residual plots; interpret over- and underestimates
| Term 术语 | Meaning 含义 | Chinese 中文 |
|---|---|---|
| Exponential Function | $f(x)=ab^x$, where $b>0,b\neq1$ | 指数函数 |
| Initial Value | $a=f(0)$, the y-intercept | 初始值 |
| Base | $b$ — the constant growth/decay factor per unit | 底数 |
| Exponential Growth | $b>1$, function increases | 指数增长 |
| Exponential Decay | $0<b<1$, function decreases | 指数衰减 |
| Horizontal Asymptote | Line $y=0$ (or $y=k$ after shift) | 水平渐近线 |
| $e$ | Euler's number $\approx 2.718$, base for continuous growth | 自然常数 |
| Residual | Observed value − Predicted value | 残差 |
| $R^2$ | Coefficient of determination (0 to 1). Not how AP validates a model — use the residual plot | 决定系数(仅作参考) |
5.1 Exponential Functions — Definition & Properties 指数函数定义与性质
若 $b\leq0$,则对某些实数 $x$(如 $x=\frac{1}{2}$)无意义(例如 $\sqrt{-2}$ 不是实数)。
所以必须 $b>0$ 且 $b\neq1$。
Properties of $f(x)=b^x$ (parent function, $a=1$)
Right end: \(x\to+\infty,\ y\to+\infty\)
Right end: \(x\to+\infty,\ y\to0^+\)
For each function, state whether it represents exponential growth or decay. Identify the initial value, base, and horizontal asymptote.
- $f(x) = 3\cdot 2^x$
- $g(x) = 5\cdot\left(\dfrac{1}{3}\right)^x$
- $h(x) = 0.4\cdot(1.25)^x$
Initial value $a=3$; Base $b=2>1$ → Exponential Growth 指数增长
HA: $y=0$; y-intercept: $(0,3)$; Domain: $\mathbb{R}$; Range: $(0,+\infty)$
Initial value $a=5$; Base $b=\frac{1}{3}<1$ → Exponential Decay 指数衰减
HA: $y=0$; y-intercept: $(0,5)$
Initial value $a=0.4$; Base $b=1.25>1$ → Exponential Growth 指数增长
HA: $y=0$; y-intercept: $(0,0.4)$
Classify each as growth or decay, and state the y-intercept and HA:
(a) $p(x)=2\cdot(0.8)^x$ (b) $q(x)=\dfrac{1}{4}\cdot 3^x$ (c) $r(x)=6\cdot e^{-x}$
(b) $b=3>1$ → Growth; y-int $(0,\frac{1}{4})$; HA $y=0$
(c) $e^{-x}=\left(\frac{1}{e}\right)^x$, $b=e^{-1}\approx0.368<1$ → Decay; y-int $(0,6)$; HA $y=0$
5.2 End Behavior of Exponential Functions 端点行为
$\displaystyle\lim_{x\to-\infty}b^x=0$
右端趋向 $+\infty$,左端趋向 $0$(从上方趋近水平渐近线)
$\displaystyle\lim_{x\to-\infty}b^x=+\infty$
右端趋向 $0$,左端趋向 $+\infty$
水平渐近线仍是 $y=0$,但图像在x轴下方。
Use the properties of $f(x)=2^x$, $g(x)=\left(\dfrac{1}{2}\right)^x$, and $h(x)=-3^x$ to evaluate:
- $\displaystyle\lim_{x\to+\infty}f(x)$ and $\displaystyle\lim_{x\to-\infty}f(x)$
- $\displaystyle\lim_{x\to+\infty}g(x)$ and $\displaystyle\lim_{x\to-\infty}g(x)$
- $\displaystyle\lim_{x\to+\infty}h(x)$ and $\displaystyle\lim_{x\to-\infty}h(x)$
$\displaystyle\lim_{x\to+\infty}2^x=+\infty$ $\displaystyle\lim_{x\to-\infty}2^x=0$
$\displaystyle\lim_{x\to+\infty}\left(\tfrac{1}{2}\right)^x=0$ $\displaystyle\lim_{x\to-\infty}\left(\tfrac{1}{2}\right)^x=+\infty$
Note: $\left(\frac{1}{2}\right)^x = 2^{-x}$, so it mirrors $2^x$ reflected over the y-axis.
$\displaystyle\lim_{x\to+\infty}(-3^x)=-\infty$ $\displaystyle\lim_{x\to-\infty}(-3^x)=0$
HA still $y=0$, but approached from below.
5.3 Laws of Exponents 指数法则
| Law 法则 | Formula 公式 | Example 例子 |
|---|---|---|
| Multiplication 乘法 | $b^m\cdot b^n = b^{m+n}$ | $2^3\cdot2^4=2^7=128$ |
| Division 除法 | $\dfrac{b^m}{b^n}=b^{m-n}$ | $\dfrac{3^5}{3^2}=3^3=27$ |
| Power of a Power 幂的乘方 | $(b^m)^n=b^{mn}$ | $(2^3)^4=2^{12}$ |
| Negative Exponent 负指数 | $b^{-n}=\dfrac{1}{b^n}$ and $\dfrac{1}{b^{-n}}=b^n$ | $5^{-2}=\dfrac{1}{25}$ |
| Rational Exponent 有理指数 | $b^{m/n}=\sqrt[n]{b^m}=\left(\sqrt[n]{b}\right)^m$ | $8^{2/3}=\left(\sqrt[3]{8}\right)^2=4$ |
| Zero Exponent 零指数 | $b^0=1$ ($b\neq0$) | $100^0=1$ |
Simplify each expression:
- $\dfrac{3^8\cdot 3^{-2}}{3^3}$
- $\left(4x^3\right)^2\cdot x^{-4}$
- $\dfrac{(2^3)^2}{2^4\cdot 2}$
- $27^{2/3}$
Simplify: (a) $\left(\dfrac{x^3}{y^2}\right)^{-2}$ (b) $\dfrac{5^{n+2}}{5^n}$ (c) $\left(4^{1/2}\cdot 8^{2/3}\right)^3$
(b) $\dfrac{5^{n+2}}{5^n}=5^2=25$
(c) $4^{1/2}=2$, $8^{2/3}=4$ → $(2\cdot4)^3=8^3=512$
5.4 Transformations & Equivalences 变换与等价形式
Product Property → Horizontal Shift = Vertical Dilation
水平平移 $h$ 单位 ≡ 垂直乘以 $b^h$
等价于将 $2^x$ 垂直拉伸 8 倍。
Power Property → Horizontal Dilation = Change of Base
水平伸缩 ≡ 改变底数
Graph $f(x)=2^x$ and $g(x)=2^{x+3}$ on the same axes. State the transformation from $f$ to $g$.
$g(x)=2^{x+3}=f(x+3)$ → horizontal shift left by 3.
Every point $(x,y)$ on $f$ maps to $(x-3,y)$ on $g$.
$f$: $(-2,\frac{1}{4}),(0,1),(1,2),(2,4)$
$g$: $(-5,\frac{1}{4}),(-3,1),(-2,2),(-1,4)$
$g(x)=2^{x+3}=2^3\cdot2^x=8\cdot2^x$
Also: vertical stretch by factor 8 of $f(x)$. HA still $y=0$.
Given $f(x)=5^{x-1}$:
(a) State its parent function.
(b) Describe the transformation from parent to $f$.
(c) Write an equivalent function $g(x)$ showing it as a vertical dilation.
(d) Describe the transformation from $g(x)$ to its parent.
$p(x)=5^x$
$f(x)=5^{x-1}=p(x-1)$ → horizontal shift right 1 向右平移1单位
$f(x)=5^{x-1}=5^{-1}\cdot5^x=\dfrac{1}{5}\cdot5^x$
So $g(x)=\dfrac{1}{5}\cdot5^x$ — vertical compression by $\frac{1}{5}$ 垂直压缩为 $\frac{1}{5}$ 倍
$g(x)=\frac{1}{5}\cdot5^x$ → vertical shrink by factor $\frac{1}{5}$
Given $f(x)=81^x$ and parent $p(x)=3^x$, find the factor of the horizontal dilation that maps $p$ to $f$.
$81=3^4$ so $81^x=(3^4)^x=3^{4x}=p(4x)$
$f(x)=p(4x)$ → horizontal compression by factor $\dfrac{1}{4}$
水平压缩为原来的 $\dfrac{1}{4}$(即 $b=4$ 意味着图像变窄到原来的 $\frac{1}{4}$)
Show that $f(x)=\left(\dfrac{1}{4}\right)^x$ is a reflection of $g(x)=4^x$ over the y-axis.
$\left(\dfrac{1}{4}\right)^x = (4^{-1})^x = 4^{-x} = g(-x)$
Since $f(x)=g(-x)$, the graph of $f$ is the graph of $g$ reflected over the y-axis. ✓
关于y轴的反射等价于将底数取倒数:$b\leftrightarrow\frac{1}{b}$
Given $f(x)=3^x$, write the equation for each transformation:
(a) Shift right 2, stretch vertically by 4
(b) Reflect over x-axis, shift up 5
(c) Horizontal compression by $\frac{1}{3}$ (equivalent base change)
(b) $-3^x+5$; HA at $y=5$
(c) Compression by $\frac{1}{3}$ → replace $x$ with $3x$: $3^{3x}=(3^3)^x=27^x$ — base changes to 27
5.5 Exponential Growth & Decay Models 指数增长与衰减模型
$a$=初始量, $r$=年增长/衰减率, $t$=时间(年)
$a$=初始量, $k$=增长率常数, $e\approx2.718$
$k>0$: growth; $k<0$: decay
其中 $P$ 为本金,$r$ 为年利率,$t$ 为年数。
半衰期 Half-life:当量减少到初始值的一半时所经过的时间。
若半衰期为 $T_{1/2}$,则模型为 $A(t)=a\cdot\left(\dfrac{1}{2}\right)^{t/T_{1/2}}$
Determine whether each table represents a linear, exponential, or neither model:
| Table A | Table B | |||
|---|---|---|---|---|
| $x$ | $y$ | $x$ | $y$ | |
| 0 | 2 | 0 | 3 | |
| 1 | 6 | 1 | 6 | |
| 2 | 18 | 2 | 9 | |
| 3 | 54 | 3 | 12 | |
$\dfrac{6}{2}=3,\quad\dfrac{18}{6}=3,\quad\dfrac{54}{18}=3$ → constant ratio = 3
✓ Exponential: $y=2\cdot3^x$
$6-3=3,\quad9-6=3,\quad12-9=3$ → constant difference = 3
✓ Linear: $y=3x+3$
连续项之比恒定 → 指数函数
Determine an equation for the exponential function whose graph passes through $(0,\,4)$ and $(5,\,8.05)$.
Point $(0,4)$: $a\cdot b^0=a=4$ → $a=4$
$4\cdot b^5=8.05$
$b^5=\dfrac{8.05}{4}=2.0125$
$b=2.0125^{1/5}\approx1.15$
$$f(x)=4\cdot(1.15)^x$$ Base $b=1.15>1$ → exponential growth at 15% per unit 指数增长,增长率15%
$\$450$ is deposited in a bank account earning 8.2% interest compounded annually. Find the balance after:
(a) 1 year (b) 2 years (c) $t$ years
年利率 8.2%,使用离散增长公式。
$A(1)=450\times1.082=\$486.90$
$A(2)=450\times(1.082)^2=450\times1.170724\approx\$526.83$
$$A(t)=450(1.082)^t$$
A culture of 100 bacteria doubles every hour. Write a model for total bacteria $B(t)$ after $t$ hours. Predict when the count will reach 5,000.
Initial value: 100; base: 2 (doubles each hour)
$$B(t)=100\cdot2^t$$
$100\cdot2^t=5000$
$2^t=50$
$t=\log_2 50=\dfrac{\ln50}{\ln2}\approx\dfrac{3.912}{0.693}\approx5.64$ hours
约 5.64小时后细菌数量达到5000。
对数(log)在第 7 章才系统讲。现在也可以用计算器:画 $Y_1=100\cdot2^X$ 与 $Y_2=5000$,用 2nd TRACE → 5: intersect 求交点,得 $t\approx5.644$。
The half-life of a radioactive substance is 20 days, and there are 5 grams initially. Write a function modeling this situation. Predict when 1 gram will remain.
$A(t)=5\cdot\left(\dfrac{1}{2}\right)^{t/20}$
每20天减少一半,所以 $t/20$ 是经过的半衰期数量。
$5\cdot\left(\dfrac{1}{2}\right)^{t/20}=1$
$\left(\dfrac{1}{2}\right)^{t/20}=\dfrac{1}{5}$
$\dfrac{t}{20}\cdot\ln\dfrac{1}{2}=\ln\dfrac{1}{5}$
$\dfrac{t}{20}=\dfrac{\ln(1/5)}{\ln(1/2)}=\dfrac{\ln5}{\ln2}\approx\dfrac{1.609}{0.693}\approx2.322$
$t=20\cdot\dfrac{\ln5}{\ln2}\approx\boxed{46.439 \text{ days}}$
不要把 $2.322$ 这种中间结果提前舍入;也可以用计算器 intersect 解 $5(0.5)^{X/20}=1$。
$\$12{,}000$ is invested for 7 years at 5% annual interest, compounded continuously. What is the balance at the end of 7 years?
$A=Pe^{rt}$, where $P=12000$, $r=0.05$, $t=7$
$A=12000\cdot e^{0.05\times7}=12000\cdot e^{0.35}$
$e^{0.35}\approx1.419068$ (keep the full calculator value; do not round early)
$A\approx\boxed{\$17{,}028.81}$
Christopher buys a new car for $\$32{,}750$. The car depreciates by about 11.75% per year. Write an equation to model this situation. Determine the value of the car after 4 years.
After 4 years: $V(4)=32750\times(0.8825)^4\approx32750\times0.606539\approx\$19{,}864.16$
Determine the exponential function that has an initial value of 15 and increases at a rate of 12% each year. Then find the value after 10 years.
After 10 years: $f(10)=15\times(1.12)^{10}=15\times3.1058\approx46.59$
Model from Two Points 由两个点建模
An exponential function $f(x)=ab^x$ satisfies $f(2)=12$ and $f(5)=96$. Find $a$ and $b$, and write $f$ in point-ratio form.
$\dfrac{ab^5}{ab^2}=\dfrac{96}{12}\Rightarrow b^3=8\Rightarrow b=2$.
$a\cdot2^2=12\Rightarrow a=3$, so $f(x)=3\cdot2^x$.
$f(x)=12\cdot2^{\,x-2}$ (check: $12\cdot2^{-2}=3$ ✓).
A town had 8,400 people in 2012 and 10,200 people in 2018. Assume the population grows exponentially. Let $t$ be years since 2012.
(a) Find a model $P(t)$. (b) What is the annual percent growth rate? (c) Predict the population in 2025. (d) Would it be reasonable to use the model to predict the population in 2150? Explain.
(b) About $3.289\%$ per year.
(c) $t=13$: $P(13)\approx12{,}793$ people.
(d) No. 2150 is far outside the data ($2012$–$2018$); a town cannot grow by the same percent forever (limited land, resources). Predictions should stay close to the data's domain (CED 2.5.B.3).
Equivalent Forms: Changing the Time Unit 换时间单位的等价形式
- $f(d)=2^d$ (doubles every day) $=\left(2^7\right)^{d/7}=128^{\,d/7}$ (multiplies by 128 every week).
- Annual factor $1.06$: monthly factor $1.06^{1/12}$, decade factor $1.06^{10}$.
- Half-life form $\left(\tfrac12\right)^{t/20}=\left(0.5^{1/20}\right)^t$: factor per day.
(a) An account grows by 6% per year: $A(t)=1000(1.06)^t$, $t$ in years. Rewrite to find the growth factor and percent change per month and per decade.
(b) The half-life model $A(t)=5\left(\tfrac12\right)^{t/20}$ ($t$ in days) from T-12: by what percent does the amount decrease each day?
$t$ years $=12t$ months: $A=1000\left(1.06^{1/12}\right)^{12t}\approx1000(1.004868)^{12t}$ → about $0.487\%$ per month (not $6\%\div12=0.5\%$).
$A=1000\left(1.06^{10}\right)^{t/10}\approx1000(1.790848)^{t/10}$ → about $79.085\%$ per decade (not $60\%$).
$A=5\left(0.5^{1/20}\right)^t\approx5(0.965936)^t$ → decreases by about $3.406\%$ per day.
A patient takes 200 mg of a medicine. Every 4 hours, 30% of the amount in the body is eliminated: $A(t)=200(0.7)^{t/4}$, $t$ in hours.
(a) Find the hourly decay factor and the percent eliminated each hour. (b) Find $A(10)$. (c) Rewrite $A$ with a factor per day (24 h) and interpret. (d) What domain makes sense?
(b) $A(10)=200(0.7)^{2.5}\approx81.993$ mg.
(c) $A(t)=200\left(0.7^{6}\right)^{t/24}\approx200(0.117649)^{t/24}$: after each full day, only about $11.765\%$ remains (about $88.235\%$ eliminated).
(d) $t\ge0$, from when the dose is taken until the next dose (the model ignores new doses).
Shifted Exponential Models $y=ab^x+k$ 带水平渐近线的指数模型
A cup of coffee cools in a 20 °C room. Its temperature $T$ (°C) is recorded every 5 minutes.
| $t$ (min) | 0 | 5 | 10 | 15 | 20 |
|---|---|---|---|---|---|
| $T$ (°C) | 90 | 62 | 45.2 | 35.12 | 29.072 |
(a) Show that $T$ itself is not exponential but $T-20$ is. (b) Write a model $T(t)$. (c) Find $\displaystyle\lim_{t\to\infty}T(t)$ and interpret. (d) Predict $T(30)$. (e) When does the coffee reach 25 °C?
$T$: $\tfrac{62}{90}\approx0.689$, $\tfrac{45.2}{62}\approx0.729$ → not constant.
$T-20$: $70,\ 42,\ 25.2,\ 15.12,\ 9.072$ → ratios $\tfrac{42}{70}=\tfrac{25.2}{42}=\dots=0.6$ → constant every 5 minutes.
$T(t)=20+70(0.6)^{t/5}$ (initial difference 70, factor 0.6 per 5 min, shifted up 20).
$(0.6)^{t/5}\to0$, so $\displaystyle\lim_{t\to\infty}T(t)=20$: the coffee approaches room temperature (HA $y=20$).
| $x$ | 0 | 1 | 2 | 3 | 4 |
|---|---|---|---|---|---|
| $f(x)$ | 5 | 7 | 11 | 19 | 35 |
(a) Show that $f$ is neither linear nor exponential as given. (b) Find a constant $k$ so that $f(x)-k$ is proportional over equal intervals. (c) Write a model for $f$ and find $\displaystyle\lim_{x\to-\infty}f(x)$.
(b) The differences double, a hint of an exponential pattern. Try $k=3$: $f(x)-3=2,4,8,16,32$, ratio 2 ✓.
(c) $f(x)=2\cdot2^x+3=2^{x+1}+3$; $\displaystyle\lim_{x\to-\infty}f(x)=3$ (HA $y=3$).
Exponential Regression with Technology 指数回归
- STAT → 1: Edit: inputs in L1, outputs in L2 (outputs must be positive).
- STAT → CALC → 0: ExpReg. Xlist: L1, Ylist: L2, Store RegEQ: Y1 (VARS → Y-VARS → Function → Y1).
- The calculator gives $y=a\cdot b^x$. Report $a$ and $b$ to 3 decimal places, but use the stored Y1 for predictions.
The number of users $U$ (millions) of an app $t$ years after launch:
| $t$ | 0 | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|---|
| $U$ | 3.1 | 4.4 | 6.5 | 9.3 | 13.8 | 19.9 |
(a) Explain why an exponential model is reasonable. (b) Find the exponential regression. (c) Interpret the base. (d) Predict $U(7)$ and comment on its reliability.
5.6 Competing Function Model Validation 模型适合性验证
- Residual plot — the model is appropriate if the residuals appear without pattern.
- Context — does the model's behavior (e.g. proportional growth, a maximum, a limit) make sense?
- $R^2$ — extra information only; the CED does not use it to validate a model.
① Coefficient of Determination $R^2$ 决定系数
$R^2=1$ → perfect fit; $R^2=0$ → model explains nothing.
$R^2$ 越接近 1,说明模型解释了数据中越多的变化,但$R^2$ 高不代表模型合适:用直线拟合指数型数据时,$R^2$ 也可能大于 0.95,而残差图却呈明显弧形(见下方"模型比较")。AP 判断模型是否合适,以残差图为准。
② Residuals & Residual Plots 残差与残差图
- 若残差图中的点随机分散在水平轴($y=0$)周围 → 模型适合 ✓
- 若残差图呈现明显规律(弧形、趋势等)→ 模型不适合,应换用其他模型 ✗
Comparing Linear, Quadratic & Exponential Models 同一组数据比较三种模型
A bacteria culture is measured (thousands of cells) every hour:
| $x$ (h) | 0 | 1 | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|---|---|
| $y$ | 2.0 | 2.9 | 4.6 | 6.7 | 10.3 | 15.1 | 23.4 |
(a) Find linear, quadratic, and exponential regressions. (b) Compute residuals for the linear and exponential models and compare the residual plots. (c) Which model is most appropriate? Justify. (d) Compare the three predictions at $x=10$.
Linear: $y\approx3.368x-0.818$ | Quadratic: $y\approx0.661x^2-0.596x+2.486$ | Exponential: $y\approx1.977(1.507)^x$
| $x$ | 0 | 1 | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|---|---|
| linear | 2.818 | 0.350 | −1.318 | −2.586 | −2.354 | −0.921 | 4.011 |
| exponential | 0.023 | −0.080 | 0.109 | −0.068 | 0.100 | −0.272 | 0.233 |
Linear: U-shaped pattern ✗
Exponential: no clear pattern ✓ (note the much smaller scale)
The linear residuals are positive, then negative, then positive: a U-shaped pattern, so the linear model is not appropriate. The exponential residuals are small and scattered without pattern, and bacteria grow proportionally (ratios $\approx1.5$ per hour). The exponential model is most appropriate. (The quadratic residuals, $-0.486$ to $0.707$, are larger and less random.)
Linear $\approx32.861$, quadratic $\approx62.593$, exponential $\approx119.514$ thousand. The models agree on the data but disagree badly outside it, which is why validation matters.
For each dataset description, determine whether a linear, quadratic, or exponential model is most appropriate. Justify.
- Fuel efficiency (mpg) of a vehicle at different speeds — increases then decreases.
- U.S. population from 1900–2011.
- Heights of females based on humerus bone length.
Increases then decreases → single maximum → Quadratic model 二次函数
Parabolic shape fits this scenario.
Population grows rapidly over time, successive ratios are approximately constant → Exponential model 指数函数
$P(t)=a\cdot b^t$ where $t$ = years since 1900.
Bone length and height are roughly proportional — a relatively straight trend → Linear model 线性函数
A linear regression model is applied to a dataset. The residual plot shows a clear U-shaped (curved) pattern. What does this tell you, and what should you do?
残差图呈U形说明线性模型不适合——预测误差有规律,不是随机分布。
Action 建议: Try a quadratic or exponential model instead, and check if the new residual plot shows random scatter.
Overestimate or Underestimate? 高估还是低估
- Residual > 0: actual is above the prediction → the model underestimates. 残差为正 → 模型低估
- Residual < 0: actual is below the prediction → the model overestimates. 残差为负 → 模型高估
(a) Using the exponential model from the bacteria example, does it overestimate or underestimate at $x=5$? At $x=6$? By how much?
(b) A city models how much road salt it needs for winter. Is it better for the model to overestimate or underestimate? Explain.
(c) A pharmacist models the maximum safe dose of a medicine for a patient's weight. Should the model err toward overestimating or underestimating the safe dose?
(b) Overestimate: running out of salt during a storm is worse than having some left over.
(c) Underestimate the safe dose: giving too much medicine is dangerous, while a slightly smaller dose is safer.
Chapter 5 Summary 本章总结
Key Formulas & Rules 核心公式
| Topic 主题 | Formula/Rule 公式/规则 | Note 备注 |
|---|---|---|
| Parent exponential | $f(x)=b^x$, $b>0,b\neq1$ | Domain: $\mathbb{R}$; Range: $(0,\infty)$ |
| General form | $f(x)=a\cdot b^x$ | $a$=initial value; HA: $y=0$ |
| Horizontal shift ↔ vertical dilation | $b^{x+h}=b^h\cdot b^x$ | Shift left $h$ = multiply by $b^h$ |
| Horizontal dilation ↔ base change | $b^{cx}=(b^c)^x$ | Compress by $\frac{1}{c}$ = new base $b^c$ |
| Discrete growth/decay | $A(t)=a(1\pm r)^t$ | $+r$: growth; $-r$: decay |
| Continuous growth/decay | $A(t)=a\cdot e^{kt}$ | $k>0$: growth; $k<0$: decay |
| Half-life model | $A(t)=a\cdot\left(\frac{1}{2}\right)^{t/T_{1/2}}$ | $T_{1/2}$=半衰期 |
| Continuous compounding | $A=Pe^{rt}$ | $P$=principal 本金 |
| Two points | $b^{x_2-x_1}=\dfrac{y_2}{y_1}$; $f(x)=y_1b^{x-x_1}$ | Divide to eliminate $a$ |
| Change time unit | $b^x=\left(b^c\right)^{x/c}$ | Factor per $c$ units is $b^c$ |
| Shifted model | $y=ab^x+k$ | HA $y=k$; check ratios of $y-k$ |
| ExpReg | STAT → CALC → 0: ExpReg | 3 decimal places; beware extrapolation |
| Validation | Residual $=$ actual $-$ predicted | No pattern → appropriate; $>0$ under-, $<0$ overestimate |
- 📈 底数 $b>1$ → 指数增长;$0<b<1$ → 指数衰减
- ➡️ 水平右移 $h$ ≡ 垂直乘以 $b^{-h}$(注意方向!)
- 🔢 判断数据类型:一阶差恒定→线性;连续项比恒定→指数
- 📊 残差图随机分散 → 模型适合;呈规律 → 换模型(以残差图为准,不看 $R^2$)
- ➕ 残差为正 → 低估;残差为负 → 高估
- 🌡️ 趋近某个值 $k$(如室温)→ 先减去 $k$ 再看比值,模型 $ab^x+k$
- 🧪 半衰期:每 $T_{1/2}$ 时间减半,用 $(\frac{1}{2})^{t/T_{1/2}}$
- 💰 连续复利:$A=Pe^{rt}$,记住 $e\approx2.718$
看底数 $b$:$b>1$ 是增长,$0<b<1$ 是衰减。
等价:$b^{-x}=\left(\frac{1}{b}\right)^x$ — 指数上的负号 = 底数取倒数 = 增长变衰减(关于y轴翻转)
Chapter 5 Review 章末习题
The function $f$ is given by $f(x)=3^{(2x)}$. Which of the following describes the graph of $f$ in the $xy$-plane and gives an equivalent form of $f(x)$?
答案:(C) $f(x)=g(2x)$,其中 $g(x)=3^x$:把输入乘以 2 是水平伸缩,因子 $\tfrac12$(图像向 $y$ 轴压缩)。幂的性质:$3^{2x}=(3^2)^x=9^x$。
(A) 竖直伸缩应是乘在外面:$2\cdot3^x$;(B)、(D) $3^{2x}\ne6^x$——指数里的 2 不能乘到底数上,要写成 $3^2$。
Which of the following statements is true about the function $k$ given by $k(x)=-2\left(0.5\right)^x+3$?
答案:(B) $(0.5)^x$ 递减、上凹;乘以 $-2$ 后上下翻转:变成递增、下凹。加 3 只是上移,不改变单调性和凹凸性。
检查:$k(0)=1$,$k(1)=2$,$k(2)=2.5$——在增加,但每次增加的量 $1,\ 0.5$ 越来越小(下凹),并且趋近 $y=3$。
The figure shows the graph of an exponential function $f$ of the form $f(x)=ab^x$. The coordinates of two points on the graph are labeled. What is the $y$-coordinate of the point on the graph where $x=0$?
答案:(C) 输入每增加 2,输出乘 $\dfrac{12}{48}=\dfrac14$,所以 $b^2=\dfrac14$,$b=\dfrac12$。从 $x=1$ 退回到 $x=0$ 要除以 $b$:$a=48\div\dfrac12=96$。
(B) 66 是按直线(等差)外推;(D) 192 是把 $\dfrac14$ 当成了每增加 1 的倍数(忘了开方)。
A radioactive substance has a half-life of $8$ days. A sample contains $50$ grams of the substance at time $t=0$. Which of the following is an appropriate model for $A(t)$, the number of grams of the substance remaining after $t$ days?
答案:(D) 每过 8 天乘 $\tfrac12$,所以 $t$ 天里乘了 $\tfrac t8$ 次:$A(t)=50\left(\tfrac12\right)^{t/8}$。检查:$A(8)=25$,$A(16)=12.5$ ✓。
(A) 线性模型(每天减少的量不变),而衰变每天减少的是固定的比例;(B) 把"8 天减半"当成"每天减少 8%";(C) $A(1)=50\left(\tfrac12\right)^8$,一天就几乎衰变完了。
A regression model $f$ is constructed for a data set. The residual at $x=2$ is $-0.8$, and the residual at $x=5$ is $2.1$. Which of the following statements is true?
答案:(A) 残差 = 实际值 − 预测值。$x=2$:残差为负 → 实际值比预测值小 → 模型高估。$x=5$:残差为正 → 模型低估。误差大小看残差的绝对值:$|2.1|>|-0.8|$,$x=5$ 处误差更大。
| $t$ (months) | $0$ | $1$ | $2$ | $3$ | $4$ | $5$ |
|---|---|---|---|---|---|---|
| $U(t)$ (thousands) | $2.0$ | $2.9$ | $4.4$ | $6.3$ | $9.8$ | $14.1$ |
The table gives the number of users of a new app, in thousands, $t$ months after it was launched. An exponential regression is used to model these data. What is the number of users, in thousands, at $t=7$ predicted by the exponential regression model?
答案:(D) STAT → CALC → 0: ExpReg:$U(t)\approx1.982(1.482)^t$,代入 $t=7$ 得 $\approx31.169$ 千人(用完整系数计算)。
(A) 是线性回归,(B) 是二次回归,(C) 只用了首尾两点 $(0,2)$、$(5,14.1)$ 建模——不是回归。
The population of a town is modeled by $P(t)=25{,}000(1.032)^t$, where $t$ is the number of years since 2020. According to the model, how many years does it take for the population to increase from $40{,}000$ to $60{,}000$?
答案:(A) 用计算器求 $P(t)=40000$ 和 $P(t)=60000$ 的解:$t_1\approx14.921$,$t_2\approx27.794$,相差 $\approx12.872$ 年。
更快的想法:从 40000 到 60000 是乘 $1.5$,只要解 $(1.032)^t=1.5$,与起点无关。(B)、(D) 是 $t_1$、$t_2$ 本身;(C) 把每年的增长量当成不变的 $0.032\times40000$。
The temperature of a cup of tea, in degrees Celsius, $t$ minutes after it is poured is modeled by $T(t)=a(0.92)^t+b$, where $a$ and $b$ are constants. If $T(0)=90$ and $\displaystyle\lim_{t\to\infty}T(t)=20$, for what value of $t$ is $T(t)=40$?
答案:(C) $t\to\infty$ 时 $(0.92)^t\to0$,所以 $b=20$(室温);$T(0)=a+b=90$,所以 $a=70$。解 $70(0.92)^t+20=40$:用计算器求交点,$t\approx15.024$ 分钟。
(A) 漏了 $+20$:解的是 $70(0.92)^t=40$;(B)、(D) 把 $a$ 当成了 90。
A city began tracking its public electric-vehicle charging stations in 2018. In 2019 $(t=1)$ there were $84$ stations, and in 2024 $(t=6)$ there were $210$ stations. The number of stations can be modeled by $P(t)=ab^t$, where $t$ is the number of years since 2018. A calculator is allowed.
(a) (i) Use the given data to write two equations that can be used to find the values of $a$ and $b$.
(ii) Find the values of $a$ and $b$ as decimal approximations.
(b) (i) Find the average rate of change of the number of stations, in stations per year, from $t=1$ to $t=6$.
(ii) Use the average rate of change from (i) to estimate the number of stations at $t=4$.
(iii) Is the estimate in (ii) greater than or less than the value predicted by the model $P$? Explain your reasoning using characteristics of the model.
(c) The model $P$ can be rewritten as $P=a\cdot c^{m}$, where $m$ is the number of months since 2018. Find the value of $c$, and interpret it as a percent change in context.
(a) (i) $ab=84$,$ab^6=210$。
(ii) 两式相除:$b^5=\dfrac{210}{84}=2.5$,$b=2.5^{1/5}\approx1.201$;$a=\dfrac{84}{b}\approx69.934$。
(b) (i) $\dfrac{210-84}{6-1}=25.2$ 个/年。
(ii) 从 $(1,84)$ 沿平均变化率走 3 年:$84+25.2\times3=159.6$ 个。
(iii) 模型 $P(4)=ab^4\approx145.560$,估计值大于模型值。理由:$a>0$、$b>1$,$P$ 递增且图像上凹,所以在 $t=1$ 到 $t=6$ 之间,过这两点的割线在图像上方,用割线估计会偏大。
(c) 1 年 = 12 个月:$P=ab^t=ab^{m/12}=a\left(b^{1/12}\right)^m$,所以 $c=b^{1/12}\approx1.0154$。含义:充电站数量每个月大约增长 $1.54\%$。
| $x$ (days) | $0$ | $1$ | $2$ | $3$ | $4$ | $5$ |
|---|---|---|---|---|---|---|
| $y$ (cm²) | $3.1$ | $4.4$ | $6.5$ | $9.4$ | $13.6$ | $20.1$ |
The table gives the area $y$, in square centimeters, of a mold colony $x$ days after a scientist began observing it. A linear regression gives $g(x)=3.3x+1.267$, and an exponential regression gives $f(x)=3.066(1.454)^x$. A calculator is allowed.
(a) Using the table, explain why an exponential model is more appropriate for these data than a linear model.
(b) Find the residual of the exponential model $f$ at $x=4$. Does $f$ overestimate or underestimate the area at $x=4$?
(c) The residual plots for the two models are shown. Which residual plot belongs to the linear model? Based on the residual plots, explain why $f$ is the more appropriate model.
(d) Use $f$ to predict the area at $x=8$. Give one reason this prediction is less reliable than a prediction at $x=2.5$.
(a) 输入每次增加 1:一阶差 $1.3,\ 2.1,\ 2.9,\ 4.2,\ 6.5$ 越来越大,不是常数,所以不是线性;相邻比值 $\dfrac{4.4}{3.1}\approx1.42$,$\dfrac{6.5}{4.4}\approx1.48$,$\dfrac{9.4}{6.5}\approx1.45$,$\dfrac{13.6}{9.4}\approx1.45$,$\dfrac{20.1}{13.6}\approx1.48$,大致相等——等距输入对应大致恒定的比值,是指数增长的特征。
(b) $f(4)=3.066(1.454)^4\approx13.703$,残差 $=13.6-13.703\approx-0.103$。残差为负,$f$ 高估了面积。
(c) Residual Plot II 是线性模型的:残差先正、再连续为负、最后又为正,呈明显的 U 形规律,说明直线没有抓住数据的弯曲趋势。Residual Plot I(指数模型)的残差正负随机分布、没有规律,而且都很小(绝对值不到 0.2),所以 $f$ 更合适。
(d) $f(8)=3.066(1.454)^8\approx61.25$ cm²。$x=8$ 在数据范围 $0\le x\le5$ 之外,是外推:没有证据说明霉菌在第 5 天以后还按同样的比例增长(空间、养分有限),而 $x=2.5$ 在数据范围内(内插)。
A patient takes a $400$-milligram dose of a medicine. The amount of the medicine in the patient's body $t$ hours later is modeled by $D(t)=400(0.85)^t$. A calculator is allowed.
(a) By what percent does the amount of medicine decrease each hour? Find the half-life of the medicine, to the nearest hundredth of an hour.
(b) The patient takes a second $400$-milligram dose at $t=6$. For $t\ge6$, the total amount in the body is $T(t)=400(0.85)^t+400(0.85)^{t-6}$. Show that $T(t)$ can be written in the form $k(0.85)^t$, and find the value of $k$.
(c) Find the total amount of medicine in the body at $t=8$.
(d) The medicine is no longer effective when the total amount falls below $100$ milligrams. After the second dose, at what time $t$ does this first happen?
(a) 增长因子 $0.85=1-0.15$,每小时减少 $15\%$。半衰期:解 $(0.85)^t=0.5$,用计算器求交点,$t\approx4.27$ 小时。
(b) 用积的性质 $(0.85)^{t-6}=(0.85)^t\cdot(0.85)^{-6}$:$T(t)=400(0.85)^t+400(0.85)^{-6}(0.85)^t=\big(400+400(0.85)^{-6}\big)(0.85)^t$,所以 $k=400\big(1+(0.85)^{-6}\big)\approx1460.587$。(水平平移 6 等价于竖直伸缩 $(0.85)^{-6}$ 倍。)
(c) $T(8)=k(0.85)^8\approx398.0$ mg。(也可以直接算 $400(0.85)^8+400(0.85)^2\approx109.0+289.0$。)
(d) 解 $1460.587(0.85)^t=100$:$t\approx16.50$ 小时,也就是第二次服药后约 10.5 小时。
Directions: Solve each part without a calculator. Show the work that leads to your answers.
(a) Solve $4^{x+1}=8^{x-1}$ for $x$.
(b) Rewrite $g(x)=3\cdot2^{(2x-1)}$ in the form $g(x)=ab^x$, where $a$ and $b$ are constants.
(c) The graph of $p(x)=16\cdot2^x$ is a vertical dilation of the graph of $y=2^x$. The graph of $p$ is also a horizontal translation of the graph of $y=2^x$. Describe this translation, and write $p(x)$ in the form $2^{(x-h)}$.
(d) The function $h$ is given by $h(x)=5\cdot2^x-3$. Solve $h(x)=37$, find $\displaystyle\lim_{x\to-\infty}h(x)$, and state the range of $h$.
(a) 化成同底:$4^{x+1}=2^{2x+2}$,$8^{x-1}=2^{3x-3}$。指数相等:$2x+2=3x-3$,$x=5$。(检验:$4^6=4096=8^4$ ✓)
(b) $3\cdot2^{2x-1}=3\cdot2^{2x}\cdot2^{-1}=\dfrac32\cdot(2^2)^x=1.5\cdot4^x$,即 $a=1.5$,$b=4$。
(c) $16=2^4$,所以 $p(x)=2^4\cdot2^x=2^{x+4}=2^{(x-(-4))}$:向左平移 4 个单位,$h=-4$。
(d) $5\cdot2^x-3=37\Rightarrow2^x=8\Rightarrow x=3$。$x\to-\infty$ 时 $2^x\to0$,所以 $\displaystyle\lim_{x\to-\infty}h(x)=-3$(水平渐近线 $y=-3$)。$5\cdot2^x>0$,值域 $(-3,\infty)$。