Infinite totals, and a second variable
The problems follow the sections of the theory. Those that are checked at once get new numbers each time you press New variant: repeat them until the method runs by itself. Those with a solution are for paper. Write the whole solution before you open it, then compare line by line. Last come problems in the form of the exam, with points and a marking guide.
Typing an answer. Write powers with ^ or **, a square root as sqrt(2) and a logarithm as ln(3); a fraction such as 1/72 is better than a rounded decimal. Two endpoints or two values are separated by a semicolon: -4; 5. The preview under the box shows how your answer was read. Look at it before you press Check.
Substitution: the last integration rule
Read the theory: 1. Substitution: the last integration rule →(a) Compute $\displaystyle\int_{1}^{e}\frac{(\ln x)^{2}}{x}\,dx.$
(b) Find $\displaystyle\int x\sqrt{x^{2}+9}\,dx,$ and check your answer by differentiating it.
(c) Compute $\displaystyle\int_{0}^{4}x\sqrt{x^{2}+9}\,dx$ in two ways: with the limits changed to $u,$ and with the answer of (b).
Hint
In (a) the inside function is $\ln x,$ and its derivative $1/x$ is there. In (b) put $u=x^{2}+9;$ then $x\,dx=\tfrac12du$ (Method 8.2).
(a) Substitute and change the limits. Put $u=\ln x,$ so $du=\dfrac1x\,dx.$ At $x=1,$ $u=\ln1=0;$ at $x=e,$ $u=\ln e=1.$ $$ \begin{aligned} \int_{1}^{e}\frac{(\ln x)^{2}}{x}\,dx &=\int_{0}^{1}u^{2}\,du\\ &=\Bigl[\tfrac13u^{3}\Bigr]_{0}^{1}=\tfrac13\,. \end{aligned} $$
(b) The antiderivative. With $u=x^{2}+9,$ $du=2x\,dx,$ so $x\,dx=\tfrac12du$ and $$ \begin{aligned} &\int x\sqrt{x^{2}+9}\,dx\\ &=\tfrac12\int u^{1/2}\,du\\ &=\tfrac12\cdot\tfrac23u^{3/2}+C\\ &=\tfrac13\bigl(x^{2}+9\bigr)^{3/2}+C\,. \end{aligned} $$
(b) The check. By the chain rule, $$ \begin{aligned} &\tfrac13\cdot\tfrac32\bigl(x^{2}+9\bigr)^{1/2}\cdot2x\\ &=x\sqrt{x^{2}+9}\,. \end{aligned} $$
(c) With the limits in $u.$ $x=0$ gives $u=9$ and $x=4$ gives $u=25{:}$ $$ \begin{aligned} &\tfrac12\int_{9}^{25}u^{1/2}\,du\\ &=\tfrac13\Bigl[u^{3/2}\Bigr]_{9}^{25}\\ &=\tfrac13(125-27)=\tfrac{98}{3}\,. \end{aligned} $$
(c) With the answer of (b), back in $x$ and with the old limits: $$ \begin{aligned} &\Bigl[\tfrac13\bigl(x^{2}+9\bigr)^{3/2}\Bigr]_{0}^{4}\\ &=\tfrac13\bigl(25^{3/2}-9^{3/2}\bigr)=\tfrac{98}{3}\,. \end{aligned} $$ The same number, as it must be. What is never allowed is the mixture: $u$ in the integrand and $0,4$ as limits.
(a) $\tfrac13.$
(b) $\tfrac13\bigl(x^{2}+9\bigr)^{3/2}+C.$
(c) $\tfrac{98}{3}\approx32.7,$ both ways.
Infinite totals
Read the theory: 2. Infinite totals →Consider the series $$ \begin{aligned} &\sum_{n=1}^{\infty}\frac{1}{n(n+1)}\\ &=\frac{1}{1\cdot2}+\frac{1}{2\cdot3}+\frac{1}{3\cdot4}+\cdots\,. \end{aligned} $$
(a) Show that $\dfrac{1}{n(n+1)}=\dfrac1n-\dfrac{1}{n+1}.$
(b) Use (a) to find the partial sum $s_{n}$ in closed form.
(c) Does the series converge? If so, what is its sum? Use Definition 8.5 and nothing else.
(d) Is the answer to (c) consistent with Fact 8.11?
Hint
In (b), write out $s_{4}$ with the terms of (a), and watch what cancels.
(a) Over the common denominator, $$ \begin{aligned} \frac1n-\frac{1}{n+1}&=\frac{(n+1)-n}{n(n+1)}\\ &=\frac{1}{n(n+1)}\,. \end{aligned} $$
(b) Almost everything cancels. $$ \begin{aligned} s_{n}&=\Bigl(1-\tfrac12\Bigr)+\Bigl(\tfrac12-\tfrac13\Bigr)\\ &\quad+\cdots+\Bigl(\tfrac1n-\tfrac{1}{n+1}\Bigr)\\ &=1-\frac{1}{n+1}\,. \end{aligned} $$ Each $-\tfrac1j$ is cancelled by the $+\tfrac1j$ of the next bracket; only the first and the last term survive. For instance $s_{1}=\tfrac12,$ $s_{2}=\tfrac12+\tfrac16=\tfrac23,$ and $1-\tfrac13=\tfrac23.$
(c) The limit. As $n\to\infty,$ $\dfrac{1}{n+1}\to0,$ so $s_{n}\to1.$ By Definition 8.5 the series converges, and its sum is $1.$
(d) Consistent. The terms satisfy $0<\dfrac{1}{n(n+1)}<\dfrac{1}{n^{2}},$ and $\sum1/n^{2}$ converges by Fact 8.11, since $p=2>1.$ The terms shrink like a power with exponent $2,$ above the borderline $1.$
(b) $s_{n}=1-\dfrac{1}{n+1}.$
(c) It converges, with sum $1.$
(d) Yes: the terms are below $1/n^{2},$ a power above the borderline.
Consider $$\sum_{n=1}^{\infty}\bigl(x^{2}-3\bigr)^{n}\,.$$
(a) For which $x$ does the series converge? The answer is not a single interval.
(b) For those $x,$ find the sum.
(c) For which $x$ is the sum equal to $1$?
Hint
The ratio is $k=x^{2}-3$ and the first term is $k$ as well. Write $|k|<1$ as a double inequality, and solve each half for $x^{2}$ first.
(a) The ratio. The series is $k+k^{2}+\cdots$ with $k=x^{2}-3,$ a geometric series with first term $k.$ By Fact 8.7 it converges exactly when $|k|<1{:}$ $$ \begin{aligned} &-1<x^{2}-3<1\\ &\iff 2<x^{2}<4\\ &\iff \sqrt2<|x|<2\,. \end{aligned} $$
(a) Two intervals. $\sqrt2<|x|<2$ holds for $\sqrt2<x<2$ and for $-2<x<-\sqrt2.$ The set is $$(-2,-\sqrt2)\cup(\sqrt2,2)\,.$$ Between $-\sqrt2$ and $\sqrt2$ the ratio is $\le-1;$ outside $[-2,2]$ it is $>1.$
(b) The sum. $$ \begin{aligned} \frac{k}{1-k}&=\frac{x^{2}-3}{1-(x^{2}-3)}\\ &=\frac{x^{2}-3}{4-x^{2}}\,. \end{aligned} $$
(c) Sum $1.$ $\dfrac{k}{1-k}=1$ gives $k=\tfrac12,$ so $x^{2}=\tfrac72$ and $x=\pm\sqrt{7/2}\approx\pm1.87.$ Both lie in the set of (a), since $\sqrt2\approx1.41<1.87<2.$
(a) $-2<x<-\sqrt2$ or $\sqrt2<x<2.$
(b) $\dfrac{x^{2}-3}{4-x^{2}}.$
(c) $x=\pm\sqrt{7/2}\approx\pm1.87.$
Present value: perpetuities, annuities and loans
Read the theory: 3. Present value: perpetuities, annuities and loans →A loan of $300\,000$ kr is repaid over $10$ years at $4\%$ a year, in equal instalments at the end of each year.
(a) Find the instalment.
(b) How much of the first instalment is interest, and how much of it reduces the debt?
(c) How much is paid in total, and what is the difference between that and $300\,000$ kr?
Hint
Fact 8.9 gives the instalment. In the first year the interest is $4\%$ of the whole debt.
(a) The instalment. By Fact 8.9 with $L=300\,000,$ $r=0.04,$ $n=10{:}$ $$ \begin{aligned} a&=300\,000\cdot\frac{0.04}{1-1.04^{-10}}\\ &=\frac{12\,000}{1-1.04^{-10}}\\ &\approx36\,987\text{ kr.} \end{aligned} $$
(b) The first year. The interest on $300\,000$ kr for one year is $0.04\cdot300\,000=12\,000$ kr. The rest of the first instalment, about $36\,987-12\,000=24\,987$ kr, reduces the debt, which is then about $275\,013$ kr. (Each later year the interest is smaller, because the debt is, so more of the same instalment goes to the debt.)
(c) The total. $10\cdot36\,987\approx369\,873$ kr. The difference, about $69\,873$ kr, is the interest: what the lender is paid for not having the money. It is not a fee on top; the ten instalments, each discounted back to today, add up to exactly $300\,000$ kr.
(a) About $36\,987$ kr a year.
(b) $12\,000$ kr interest, about $24\,987$ kr repayment.
(c) About $369\,873$ kr; the $69\,873$ kr above the loan is interest.
A share pays a dividend of $50$ kr at the end of next year, and the dividend then grows by $2\%$ a year, for ever. The interest rate is $6\%$ a year.
(a) Write the value of the share today as a series, and show that it is geometric. What are $a$ and $k$?
(b) Find the value. Compare with a share whose dividend stays at $50$ kr.
(c) What happens to your answer if the dividend grows by $6\%$ a year or more? What does that say about the model?
Hint
The dividend at the end of year $n$ is $50\cdot1.02^{n-1}$ kr, and it is worth that divided by $1.06^{n}$ today (Example 8.8).
(a) The series. The dividend at the end of year $n$ is $50\cdot1.02^{n-1},$ worth $50\cdot1.02^{n-1}/1.06^{n}$ today: $$ \begin{aligned} V&=\frac{50}{1.06}+\frac{50\cdot1.02}{1.06^{2}}\\ &\quad+\frac{50\cdot1.02^{2}}{1.06^{3}}+\cdots\,. \end{aligned} $$ Each term is the one before times $\dfrac{1.02}{1.06},$ so it is geometric with $a=\dfrac{50}{1.06}$ and $k=\dfrac{1.02}{1.06}\approx0.962.$
(b) The value. $|k|<1,$ so by Fact 8.7 $$ \begin{aligned} V&=\frac{50/1.06}{1-1.02/1.06}\\ &=\frac{50/1.06}{0.04/1.06}\\ &=\frac{50}{0.04}=1250\text{ kr.} \end{aligned} $$ Without growth it is the perpetuity $50/0.06\approx833$ kr (Example 8.8). Two per cent of growth adds half as much again.
(c) When growth catches up. With growth $g$ the ratio is $k=(1+g)/1.06.$ If $g\ge6\%$ then $k\ge1,$ and by Fact 8.7 the series diverges: the discounted dividends do not shrink, and the “value” is infinite. No one pays an infinite price, so the model breaks: a dividend cannot grow at least as fast as the interest rate for ever. The formula $50/(r-g)$ that (b) suggests is meaningless there; it would even give a negative price for $g>r.$
(a) Geometric, with $a=50/1.06$ and $k=1.02/1.06.$
(b) $1250$ kr, against $833$ kr without growth.
(c) For growth $\ge6\%$ the ratio is $\ge1$ and the series diverges: the model gives no finite value.
A second variable
Read the theory: 4. A second variable →Let $F(K,L)=4K^{1/4}L^{1/2}.$
(a) Compute $F(16,9).$
(b) Both inputs are multiplied by $t>0.$ By what factor is output multiplied? What kind of returns to scale is this, and by about how much does output rise when both inputs double?
(c) Now only capital doubles, from $16$ to $32,$ and labour stays at $9.$ By what factor does output rise?
(d) Which exponent on $L$ would give constant returns to scale?
Hint
Substitute and collect the powers of $t,$ as in Example 8.14. In (c) only one factor changes.
(a) $16^{1/4}=2$ and $9^{1/2}=3,$ so $F(16,9)=4\cdot2\cdot3=24.$
(b) Scale both. $$ \begin{aligned} F(tK,tL)&=4\,t^{1/4}K^{1/4}\,t^{1/2}L^{1/2}\\ &=t^{3/4}F(K,L)\,. \end{aligned} $$ $\tfrac14+\tfrac12=\tfrac34<1{:}$ decreasing returns to scale. Doubling both inputs multiplies output by $2^{3/4}\approx1.68.$
(c) One input alone. Only the factor $K^{1/4}$ changes, so output is multiplied by $2^{1/4}\approx1.19{:}$ $F(32,9)=24\cdot2^{1/4}\approx28.5.$ Doubling one input does much less than doubling both.
(d) Constant returns need the exponents to add up to $1{:}$ $\tfrac14+b=1,$ so $b=\tfrac34.$
(a) $24.$
(b) By $t^{3/4};$ decreasing returns; about $1.68$ when both double.
(c) By $2^{1/4}\approx1.19.$
(d) $\tfrac34.$
Partial derivatives
Read the theory: 5. Partial derivatives →A household's demand for a good, at price $p$ and income $m,$ is $$D(p,m)=\frac{20\sqrt{m}}{p}\qquad(p>0,\ m>0)\,.$$
(a) Find $\dfrac{\partial D}{\partial p}$ and $\dfrac{\partial D}{\partial m}.$
(b) Evaluate $D$ and both partial derivatives at $(p,m)=(5,400),$ and say in words what each partial derivative means.
(c) Compute the two elasticities $\dfrac{p}{D}\cdot D'_{p}$ and $\dfrac{m}{D}\cdot D'_{m}$ at $(5,400).$ Where could you have read them off?
(d) Next year the price rises to $6$ and income to $500.$ Why is $-16$ not a good estimate of the change in demand?
Hint
Write $D=20\,m^{1/2}p^{-1}.$ Each partial derivative is the power rule in one variable, with the other a constant (Watch out 8.16).
(a) With $D=20\,m^{1/2}p^{-1}{:}$ $$ \begin{aligned} \frac{\partial D}{\partial p}&=-\frac{20\sqrt{m}}{p^{2}}\,,\\ \frac{\partial D}{\partial m}&=\frac{10}{p\sqrt{m}}\,. \end{aligned} $$
(b) At $(5,400).$ $\sqrt{400}=20,$ so $D=\dfrac{20\cdot20}{5}=80,$ $D'_{p}=-\dfrac{400}{25}=-16$ and $D'_{m}=\dfrac{10}{5\cdot20}=0.1.$
In words: with income held at $400,$ raising the price by one unit lowers demand by about $16$ units; with the price held at $5,$ one more unit of income raises demand by about $0.1$ units. Each holds the other variable fixed: ceteris paribus.
(c) The elasticities. $$ \begin{aligned} \frac{p}{D}\,D'_{p}&=\frac{5}{80}\cdot(-16)=-1\,,\\ \frac{m}{D}\,D'_{m}&=\frac{400}{80}\cdot0.1=0.5\,. \end{aligned} $$ $D=20\,m^{1/2}p^{-1}$ has the form of a Cobb–Douglas function, so the elasticities are the exponents, $-1$ and $\tfrac12,$ as after Example 8.17: a price $1\%$ higher lowers demand by about $1\%,$ an income $1\%$ higher raises it by about $\tfrac12\%.$
(d) Two objections. $D'_{p}=-16$ holds income fixed, but income rises too, which pushes demand up; and a price change of a whole unit (from $5$ to $6,$ that is $20\%$) is not small, while $D'_{p}$ changes with $p$ (Watch out 8.18). In fact $D(6,500)=\dfrac{20\sqrt{500}}{6}\approx74.5,$ a fall of about $5.5,$ not $16.$
(a) $D'_{p}=-\dfrac{20\sqrt m}{p^{2}},$ $D'_{m}=\dfrac{10}{p\sqrt m}.$
(b) $D=80,$ $D'_{p}=-16,$ $D'_{m}=0.1.$
(c) $-1$ and $0.5,$ the exponents.
(d) Income changes too, and the price change is not small; the true change is about $-5.5.$
Differentiating twice
Read the theory: 6. Differentiating twice →Let $f(x,y)=x\,e^{xy}.$
(a) Find $f'_{1}$ and $f'_{2}.$
(b) Find all four second-order partial derivatives, and check that the mixed ones agree.
(c) Evaluate the four at $(x,y)=(1,0).$
Hint
For $f'_{1},$ $x\,e^{xy}$ is a product of two factors that both contain $x{:}$ product rule and chain rule. For $f'_{2},$ the factor $x$ is a constant.
(a) First order. With $y$ fixed, by the product rule and the chain rule, $$ \begin{aligned} f'_{1}&=e^{xy}+x\cdot y\,e^{xy}\\ &=(1+xy)\,e^{xy}\,. \end{aligned} $$ With $x$ fixed, $f'_{2}=x\cdot x\,e^{xy}=x^{2}e^{xy}.$
(b) Differentiate $f'_{1}$ again. $$ \begin{aligned} f''_{11}&=y\,e^{xy}+(1+xy)\,y\,e^{xy}\\ &=(2y+xy^{2})\,e^{xy}\,,\\ f''_{12}&=x\,e^{xy}+(1+xy)\,x\,e^{xy}\\ &=(2x+x^{2}y)\,e^{xy}\,. \end{aligned} $$
(b) Differentiate $f'_{2}$ again. $$ \begin{aligned} f''_{21}&=2x\,e^{xy}+x^{2}y\,e^{xy}\\ &=(2x+x^{2}y)\,e^{xy}\,,\\ f''_{22}&=x^{3}e^{xy}\,. \end{aligned} $$ $f''_{12}=f''_{21},$ as Fact 8.21 says. The two came by different routes, so their agreement checks both $f'_{1}$ and $f'_{2}.$
(c) At $(1,0).$ $e^{0}=1,$ so $f''_{11}=0,$ $f''_{12}=f''_{21}=2$ and $f''_{22}=1.$
(a) $f'_{1}=(1+xy)e^{xy},$ $f'_{2}=x^{2}e^{xy}.$
(b) $f''_{11}=(2y+xy^{2})e^{xy},$ $f''_{12}=f''_{21}=(2x+x^{2}y)e^{xy},$ $f''_{22}=x^{3}e^{xy}.$
(c) $0,$ $2,$ $2,$ $1.$
Mixed, as on the exam
Here the sections are mixed, and nothing says which method to use. Write the solution on paper, with reasons, before you look.
Let $f(x)=e^{-x/2}.$
(a) Compute $\displaystyle\int_{0}^{\infty}e^{-x/2}\,dx.$
(b) Compute $\displaystyle\sum_{n=0}^{\infty}e^{-n/2}$ and $\displaystyle\sum_{n=1}^{\infty}e^{-n/2},$ exactly and to two decimals.
(c) Explain with rectangles of width $1$ why $$\sum_{n=1}^{\infty}e^{-n/2}<\int_{0}^{\infty}e^{-x/2}\,dx<\sum_{n=0}^{\infty}e^{-n/2}\,.$$
Hint
In (a) integrate to $b$ and let $b\to\infty$ (Definition 8.5). In (b), $e^{-n/2}=\bigl(e^{-1/2}\bigr)^{n}{:}$ a geometric series. In (c), $f$ is decreasing.
(a) The integral. An antiderivative of $e^{-x/2}$ is $-2e^{-x/2}$ (check by differentiating). So $$ \begin{aligned} &\int_{0}^{b}e^{-x/2}\,dx\\ &=\Bigl[-2e^{-x/2}\Bigr]_{0}^{b}\\ &=2-2e^{-b/2}\;\longrightarrow\;2\,, \end{aligned} $$ since $e^{-b/2}\to0$ as $b\to\infty.$ The integral converges, to $2.$
(b) The sums. $e^{-n/2}=k^{n}$ with $k=e^{-1/2}\approx0.607<1.$ By Fact 8.7: $$ \begin{aligned} \sum_{n=0}^{\infty}e^{-n/2}&=\frac{1}{1-e^{-1/2}}\approx2.54\,,\\ \sum_{n=1}^{\infty}e^{-n/2}&=\frac{e^{-1/2}}{1-e^{-1/2}}\approx1.54\,. \end{aligned} $$ The second has first term $e^{-1/2},$ not $1,$ and is the first minus $1.$
(c) Rectangles. On $[n,n+1]$ the function decreases from $f(n)$ to $f(n+1).$ A rectangle of width $1$ and height $f(n)$ lies above the graph there, and one of height $f(n+1)$ lies below it. Adding over $n=0,1,2,\dots$: the tall rectangles have total area $\sum_{n\ge0}f(n)\approx2.54,$ the short ones $\sum_{n\ge1}f(n)\approx1.54,$ and the area under the graph, $2,$ lies strictly between. This is why the integral test of Fact 8.11 and the sum test have the same borderline.
(a) $2.$
(b) $\dfrac{1}{1-e^{-1/2}}\approx2.54$ and $\dfrac{e^{-1/2}}{1-e^{-1/2}}\approx1.54.$
(c) The rectangles of height $f(n+1)$ lie under the graph and those of height $f(n)$ over it: $1.54<2<2.54.$
As on the exam
4 problems · 36 pProblems in exam form, with points. Solve them on paper or in the box. Then open the solution one step at a time and mark yourself against the rubric.
Two improper integrals by substitution
Calculate the integrals
- (a)5 p$\displaystyle\int_{0}^{\infty}(x+1)\,3^{-x^{2}-2x}\,dx,$
- (b)5 p$\displaystyle\int_{0}^{\infty}\frac{4x+2}{\bigl(x^{2}+x+2\bigr)^{3}}\,dx.$
The text is stored only in this browser. Nobody marks it for you.
(a) The substitution. The exponent $-x^{2}-2x$ has derivative $-(2x+2)=-2(x+1),$ and $x+1$ stands in front: the signal of Method 8.2. Put $u=x^{2}+2x,$ so $du=(2x+2)\,dx$ and $$(x+1)\,dx=\tfrac12\,du\,.$$
(a) A finite upper limit, and the limits in $u.$ By Definition 8.5 the integral is the limit as $s\to\infty$ of the integral from $0$ to $s.$ At $x=0,$ $u=0;$ at $x=s,$ $u=s^{2}+2s.$ So $$ \begin{aligned} &\int_{0}^{s}(x+1)\,3^{-x^{2}-2x}\,dx\\ &=\tfrac12\int_{0}^{s^{2}+2s}3^{-u}\,du\,. \end{aligned} $$
(a) An antiderivative of $3^{-u}.$ By Fact 4.4 and the chain rule, the derivative of $3^{-u}$ is $-3^{-u}\ln3,$ so $-\dfrac{3^{-u}}{\ln3}$ is an antiderivative of $3^{-u}.$ Hence $$ \begin{aligned} &\tfrac12\int_{0}^{s^{2}+2s}3^{-u}\,du\\ &=\tfrac12\Bigl[-\frac{3^{-u}}{\ln3}\Bigr]_{0}^{s^{2}+2s}\\ &=\frac{1}{2\ln3}\Bigl(1-3^{-(s^{2}+2s)}\Bigr)\,. \end{aligned} $$
(a) The limit. As $s\to\infty,$ $s^{2}+2s\to\infty,$ so $3^{-(s^{2}+2s)}=1/3^{s^{2}+2s}\to0.$ The limit exists and is finite: the integral converges, and $$ \begin{aligned} &\int_{0}^{\infty}(x+1)\,3^{-x^{2}-2x}\,dx\\ &=\frac{1}{2\ln3}\approx0.455\,. \end{aligned} $$
(b) The substitution. The base of the power is $x^{2}+x+2,$ whose derivative $2x+1$ is half of the numerator. Put $u=x^{2}+x+2,$ $du=(2x+1)\,dx,$ so $(4x+2)\,dx=2\,du.$ At $x=0,$ $u=2;$ at $x=s,$ $u=s^{2}+s+2.$ (The denominator is never $0{:}$ $x^{2}+x+2\ge2$ for $x\ge0.$)
(b) The integral to $s.$ $$ \begin{aligned} &\int_{0}^{s}\frac{4x+2}{(x^{2}+x+2)^{3}}\,dx\\ &=\int_{2}^{s^{2}+s+2}2u^{-3}\,du\\ &=\Bigl[-u^{-2}\Bigr]_{2}^{s^{2}+s+2}\\ &=\frac14-\frac{1}{(s^{2}+s+2)^{2}}\,. \end{aligned} $$
(b) The limit. As $s\to\infty$ the last term tends to $0,$ so the integral converges and $$ \begin{aligned} &\int_{0}^{\infty}\frac{4x+2}{(x^{2}+x+2)^{3}}\,dx\\ &=\frac14\,. \end{aligned} $$
(a) $\dfrac{1}{2\ln3}\approx0.455.$
(b) $\dfrac14.$
Tick each criterion that is really in your solution. A criterion scores only if what it asks for is there.
Common errors
- $dx$ replaced by $du$ as if they were the same: in (a) the factor $\tfrac12$ is lost and the answer is $\frac{1}{\ln3},$ twice too large; in (b) the factor $2$ is lost. In the live vote on this step only about a fifth of the answers were right.
- The limits left in $x$ after the substitution. In (b) that puts $u=0$ as the lower limit, where $u^{-2}$ is not even defined.
- In (a), $\int3^{-u}\,du$ treated by the power rule, $\frac{3^{-u+1}}{-u+1}.$ The variable is in the exponent: write $3^{-u}=e^{-u\ln3}.$
- The upper limit $\infty$ substituted directly (“$3^{-\infty}=0$”) instead of integrating to $s$ and letting $s\to\infty.$ The HT26 solutions write the limit, and so must the answer.
A series with a root, and the tail of another
Answer both parts.
- (a)5 p
For which real values of $x$ does the series $$\sum_{n=1}^{\infty}\left(\frac{\sqrt{x+6}}{4}\right)^{n}$$ converge? For which value of $x$ is the sum of the series equal to $3$?
- (b)5 p
Find the smallest natural number $N$ such that $$\sum_{n=N}^{\infty}\left(\frac23\right)^{n}<\frac{1}{100}\,.$$
The text is stored only in this browser. Nobody marks it for you.
(a) The series is geometric. Written out it is $k+k^{2}+k^{3}+\cdots$ with $$k=\frac{\sqrt{x+6}}{4}\,,$$ so the first term is $k$ (the sum starts at $n=1$) and the ratio is $k.$
(a) The domain of the root. $\sqrt{x+6}$ is defined for $x\ge-6,$ and then $k\ge0.$ So $|k|<1$ means only $k<1.$
(a) Where it converges. By Fact 8.7 the series converges exactly when $|k|<1{:}$ $$ \begin{aligned} &\frac{\sqrt{x+6}}{4}<1\\ &\iff\sqrt{x+6}<4\\ &\iff x+6<16\\ &\iff x<10\,, \end{aligned} $$ where squaring keeps the direction because both sides are non-negative. At $x=-6$ every term is $0,$ so the series converges (to $0$); at $x=10$ the ratio is $1$ and it diverges. The series converges exactly for $-6\le x<10.$
(a) The sum. For those $x$ the sum is $\dfrac{a}{1-k}=\dfrac{k}{1-k}.$ Setting it equal to $3{:}$ $$ \begin{aligned} \frac{k}{1-k}=3&\iff k=3-3k\\ &\iff k=\tfrac34\,, \end{aligned} $$ so $\sqrt{x+6}=3,$ $x+6=9$ and $x=3,$ which lies in $[-6,10).$
(b) The tail is geometric. Written out, $\sum_{n=N}^{\infty}\bigl(\tfrac23\bigr)^{n} =\bigl(\tfrac23\bigr)^{N}+\bigl(\tfrac23\bigr)^{N+1}+\cdots,$ a geometric series with first term $\bigl(\tfrac23\bigr)^{N}$ and ratio $\tfrac23.$ Since $\tfrac23<1,$ by Fact 8.7 $$ \begin{aligned} \sum_{n=N}^{\infty}\Bigl(\frac23\Bigr)^{n} &=\frac{(2/3)^{N}}{1-2/3}\\ &=3\Bigl(\frac23\Bigr)^{N}\,. \end{aligned} $$
(b) The inequality. $3\bigl(\tfrac23\bigr)^{N}<\tfrac{1}{100}$ means $\bigl(\tfrac23\bigr)^{N}<\tfrac{1}{300}.$ Take $\ln$ of both sides (Method 2.11); $\ln$ is increasing, so the direction is kept: $$N\ln\tfrac23<\ln\tfrac{1}{300}=-\ln300\,.$$
(b) Divide by a negative number. $\ln\tfrac23<0,$ so dividing by it turns the inequality: $$ \begin{aligned} N&>\frac{-\ln300}{\ln(2/3)}=\frac{\ln300}{\ln1.5}\\ &\approx\frac{5.704}{0.4055}\approx14.07\,. \end{aligned} $$
(b) The smallest $N,$ and a check. The smallest natural number above $14.07$ is $N=15.$ Check: $3\bigl(\tfrac23\bigr)^{14}\approx0.0103$ is not below $0.01,$ and $3\bigl(\tfrac23\bigr)^{15}\approx0.0069$ is.
(a) It converges exactly for $-6\le x<10;$ the sum is $3$ at $x=3.$
(b) $N=15.$
Tick each criterion that is really in your solution. A criterion scores only if what it asks for is there.
Common errors
- The domain of the root forgotten, so that $-1<\sqrt{x+6}/4<1$ is solved as $-16<x+6<16,$ giving $-22<x<10.$ A square root is never negative, and it is undefined for $x<-6.$
- The sum taken as $\frac{1}{1-k}$ although the series starts at $n=1;$ then $\frac{1}{1-k}=3$ gives $k=\tfrac23$ and $x=\tfrac{10}{9}.$ Write out the first term before using the formula.
- In (b), the tail summed as if its first term were $1{:}$ $\frac{1}{1-2/3}=3,$ which does not depend on $N$ at all.
- In (b), the inequality not turned when dividing by $\ln\tfrac23<0,$ or $N=14$ given from $N>14.07$ without a check.
A savings plan, a perpetuity and an annuity
Answer both parts.
- (a)5 p
At the end of every year you deposit $2500$ kronor in an account that pays $4\,\%$ interest per year. Let $S_{n}$ be the balance immediately after the $n$-th deposit. Show that $$S_{n}=62\,500\,\bigl(1.04^{n}-1\bigr)\,,$$ and determine the smallest $n$ for which the balance exceeds $100\,000$ kronor.
- (b)5 p
A foundation is to award a prize of $6000$ kronor at the end of every year, and its money earns $3\,\%$ interest per year.
(i) How much money must the foundation hold today if the prize is to be awarded for ever?
(ii) Suppose instead that the prize is to be awarded for $20$ years only. Write the amount needed today as a closed formula (without a sum sign), explain how it follows from the formula for a geometric sum, and evaluate it to the nearest krona.
The text is stored only in this browser. Nobody marks it for you.
(a) The sum the words describe. Just after the $n$-th deposit the first deposit has earned interest for $n-1$ years, the second for $n-2$ years, …, and the last for none. So $$ \begin{aligned} S_{n}&=2500\bigl(1.04^{n-1}+1.04^{n-2}\\ &\qquad+\cdots+1.04+1\bigr)\,. \end{aligned} $$
(a) The closed formula. A finite geometric sum with first term $2500,$ ratio $1.04$ and $n$ terms. By the finite sum in the “why” of Fact 8.7, $a\,\dfrac{1-k^{n}}{1-k},$ valid for every $k\ne1{:}$ $$ \begin{aligned} S_{n}&=2500\cdot\frac{1-1.04^{n}}{1-1.04}\\ &=\frac{2500}{0.04}\bigl(1.04^{n}-1\bigr)\\ &=62\,500\,\bigl(1.04^{n}-1\bigr)\,. \end{aligned} $$
(a) The inequality, and logarithms. $S_{n}>100\,000$ means $1.04^{n}-1>1.6,$ that is $1.04^{n}>2.6.$ $\ln$ is increasing, so (Method 2.11) $$ \begin{aligned} &n\ln1.04>\ln2.6\\ &\iff n>\frac{\ln2.6}{\ln1.04}\\ &\qquad\approx\frac{0.9555}{0.03922}\\ &\qquad\approx24.4\,, \end{aligned} $$ dividing by $\ln1.04>0$ without turning the inequality.
(a) The smallest $n,$ and a check. So $n=25.$ Check: $S_{24}\approx97\,707<100\,000$ and $S_{25}\approx104\,115>100\,000.$
(b)(i) For ever. A prize at the end of year $j$ is worth $6000\cdot1.03^{-j}$ today. The amount needed is the present value of all of them, $$\frac{6000}{1.03}+\frac{6000}{1.03^{2}}+\cdots\,,$$ a geometric series with first term $a=6000/1.03$ and ratio $k=1/1.03<1,$ so by Fact 8.7 (as in Example 8.8) $$ \begin{aligned} \mathrm{PV}&=\frac{6000/1.03}{1-1/1.03}\\ &=\frac{6000}{0.03}=200\,000\text{ kronor.} \end{aligned} $$
(b)(ii) For twenty years. Now the same series stops after $20$ terms. The finite sum $a\,\dfrac{1-k^{n}}{1-k}$ with $a=6000/1.03,$ $k=1/1.03$ and $n=20$ gives $$ \begin{aligned} \mathrm{PV}_{20}&=\frac{6000}{1.03}\cdot\frac{1-1.03^{-20}}{1-1.03^{-1}}\\ &=6000\cdot\frac{1-1.03^{-20}}{0.03}\,, \end{aligned} $$ multiplying numerator and denominator by $1.03$ (this is Fact 8.9).
(b)(ii) The value. $1.03^{-20}\approx0.553\,676,$ so $\mathrm{PV}_{20}\approx89\,265$ kronor. As a check, it is less than $200\,000;$ the difference, about $110\,735$ kronor, is what the prizes after year $20$ are worth today, $200\,000\cdot1.03^{-20}.$
(a) $S_{n}=62\,500(1.04^{n}-1);$ the balance first exceeds $100\,000$ kronor after the $25$th deposit.
(b) (i) $200\,000$ kronor. (ii) $6000\cdot\dfrac{1-1.03^{-20}}{0.03}\approx89\,265$ kronor.
Tick each criterion that is really in your solution. A criterion scores only if what it asks for is there.
Common errors
- In (a), the powers run from $1.04$ to $1.04^{n}{:}$ that would be the balance a year after the last deposit. The deposit made today has earned no interest yet.
- In (b)(i), the first term taken as $6000$ instead of $6000/1.03,$ which counts a prize today and gives $206\,000.$ In the practice hub the series that starts at $n=1$ was answered correctly only $40\%$ of the time.
- In (b)(ii), the payments compounded forward, $6000\cdot\frac{1.03^{20}-1}{0.03},$ instead of discounted back: a value in twenty years is not a value today. In the live vote about a fifth of the answers made this error for a loan.
- In (a), $n=24$ from $n>24.4,$ or no check of $S_{24}$ and $S_{25}.$
Convergence on two rays, and a logarithm as the ratio
Answer both parts.
- (a)3 p
For which real $x$ does $$\sum_{n=1}^{\infty}(x-1)^{-2n}$$ converge? For which $x$ is its sum equal to $\tfrac18$?
- (b)3 p
For which $x>0$ does $$\sum_{n=0}^{\infty}\bigl(2\ln x\bigr)^{n}$$ converge? For which $x$ is its sum equal to $2$?
The text is stored only in this browser. Nobody marks it for you.
(a) The ratio. For $x\ne1,$ $(x-1)^{-2n}=\Bigl(\dfrac{1}{(x-1)^{2}}\Bigr)^{n},$ so the series is geometric with ratio $k=\dfrac{1}{(x-1)^{2}}>0$ and first term $k.$ (At $x=1$ the terms are not defined.)
(a) Two rays. By Fact 8.7 it converges exactly when $k<1,$ that is $$ \begin{aligned} &(x-1)^{2}>1\\ &\iff|x-1|>1\\ &\iff x<0\ \text{ or }\ x>2\,. \end{aligned} $$ For $0\le x\le2,$ $x\ne1,$ the ratio is $\ge1$ and the series diverges.
(a) The sum. For $x<0$ or $x>2,$ $$ \begin{aligned} \frac{k}{1-k}&=\frac{1}{(x-1)^{2}-1}\\ &=\frac{1}{x^{2}-2x}\,. \end{aligned} $$ It equals $\tfrac18$ when $x^{2}-2x=8,$ that is $(x-4)(x+2)=0{:}$ $x=4$ or $x=-2.$ Both lie in the set of convergence.
(b) Where it converges. The ratio is $k=2\ln x$ and the first term is $k^{0}=1.$ By Fact 8.7 the series converges exactly when $|2\ln x|<1{:}$ $$ \begin{aligned} &-\tfrac12<\ln x<\tfrac12\\ &\iff e^{-1/2}<x<e^{1/2}\,, \end{aligned} $$ since $e^{t}$ is increasing. (All these $x$ are positive, as the logarithm requires.)
(b) The sum. For those $x$ the sum is $\dfrac{1}{1-2\ln x}.$ It equals $2$ when $1-2\ln x=\tfrac12,$ that is $\ln x=\tfrac14$ and $x=e^{1/4}\approx1.28,$ which lies between $e^{-1/2}\approx0.61$ and $e^{1/2}\approx1.65.$
(a) For $x<0$ and for $x>2;$ the sum is $\tfrac18$ at $x=-2$ and at $x=4.$
(b) For $e^{-1/2}<x<e^{1/2};$ the sum is $2$ at $x=e^{1/4}.$
Tick each criterion that is really in your solution. A criterion scores only if what it asks for is there.
Common errors
- In (a), the condition written as $|x-1|<1$ because “a geometric series needs something below $1$”. Here the ratio is $1/(x-1)^{2},$ and that is what must be below $1.$
- In (a), only one of the solutions of $x^{2}-2x=8,$ or a solution not checked against the set of convergence.
- In (b), the sum written $\frac{k}{1-k}$ although the series starts at $n=0,$ so that the first term is $1.$
- In (b), only $\ln x<\tfrac12$ solved, which lets every $x$ between $0$ and $e^{-1/2}$ in, where $2\ln x<-1.$