Give complete explanations of what you are doing, written in full sentences.
Solutions that have all the correct calculations and computations, but lack explanations,
will not get full marks!
Under certain circumstances a rumour spreads according to the equation
\begin{equation*}\fe{p}{t} = \frac{1}{1+ a e^{-kt}}\end{equation*}
where \(\fe{p}{t} \) is the proportion of the population that knows the rumor at
time \(t\) (in days) and \(a\) and \(k\) are positive constants.
Note that this is an example of the Logistic Function. When you get to differential
equations this equation is quite important.
Gilbert Strang
gives a good presentation on this subject.
It is a very important equation and comes up often.
See Logistic function (wiki)
The inverse function is an example of a Logit function (wiki).
a)
Find \(\lim_{t \to \infty} p(t)\). What does this mean for the rumor?
Answer\(p(\infty) = 1\)After a long period of time everyone knows the rumour!
Solution
\begin{align*}
lim_{t \to \infty}\fe{p}{t}&=\frac{1}{1+ a e^{-k\infty}}\\
\text{Since }e^{-k\infty} = 0, \{ k \in \mathbb{R} | k > 0 \}\\
\therefore lim_{t \to \infty} p(t)&= \frac{1}{1+0}\\
&=1
\end{align*}
b)
Find the rate of spread of the rumor.
Answer\(\frac{dp}{dt}=\frac{ak e^{-kt}}{(1+a e^{-kt})^2}\) Solution
\begin{align*}
u&= 1 + a e^{-kt}\\
\frac{dp}{dt}&=\frac{dp}{du}\frac{du}{dt}&\text{(Chain Rule)}\\
&=\frac{-1}{u^2}\left(-ak e^{-kt}\right)\\
&=\frac{ak e^{-kt}}{(1+a e^{-kt})^2}
\end{align*}
c)
Find the inverse function of \(p(t)\) and give an interpretation of the meaning.
Answer\(\frac{1}{k}\ln\left(\frac{at}{1-t}\right)\) Solution
\begin{align*}
p&= \frac{1}{1+ a e^{-kt}}\\
\text{Substitute p's for t's, and vice versa:}\\
t&= \frac{1}{1+ a e^{-kp}}\\
\text{Solve for p:}\\
1 + a e^{-kp}&=\frac{1}{t}\\
a e^{-kp}&=\frac{1}{t} - 1\\
e^{-kp}&=\frac{1}{at} - \frac{1}{a}\\
-kp&=\ln \left(\frac{1}{at} - \frac{1}{a}\right)\\
p&=-\frac{1}{k}\ln \left(\frac{1-t}{at}\right)\\
p&=\frac{1}{k}\ln \left(\frac{at}{1-t}\right)\\
\text{In other words the inverse is:}\\
p^{-1}&=\frac{1}{k}\ln \left(\frac{at}{1-t}\right)
\end{align*}
d)
Graph \(p\) for the case \(a=10,k=0.5\) and use your graph to estimate how
long it will take for 80% of the population to hear the rumor.
Can you also calculate this time?
Answer
\(7.38\) days.
See Sage plot of \(p(t)\) below.
Solution
Use the inverse function with a value of \(0.8\)
\begin{align*}
p^{-1}&=\frac{1}{k}\ln \left(\frac{at}{1-t}\right)\\
p^{-1}(0.8)&=7.378
\end{align*}
See Sage plot of inverse \(p(t)\) below.
Plot of \(p(t)\).
From the sage plot we can see that the time to get to 80% is about \(7.5\) days.
Plot of \(p^{-1}(t)\) (the inverse of \(p(t)\)).
From the sage calculation we can see that the time to get to 80% is \(7.37775890822787\) days.
Bismuth-210 has a half-life of 5.0 days.
a)
A sample originally has a mass of \(800 mg\).
Find a formula for the mass remaining after \(t\) days.
Answer
The mass (in \(mg\)) after time \(t\) (in days) is:
\begin{equation*}M_t=800 e^{-0.1386t}\end{equation*}
Solution
The rate of change of mass is proportional to the current mass:
\begin{equation*}\frac{dM}{dt} \propto M\end{equation*}
Lets call the constant of proportionality \(\lambda\).
\begin{equation*}\frac{dM}{dt} = -\lambda M\end{equation*}
The negative is because this is a decay.
The solution of this equation is:
\begin{equation*}M_t = M_0 e^{-\lambda t}\end{equation*}
(where \(M_0\) is the mass at \(t=0\).)
At the half-life \(t=T_{\frac{1}{2}}\) and \(M_t = \frac{M_0}{2}\):
\begin{equation*}\frac{M_0}{2} = M_0 \exp\left({-\lambda T_{\frac{1}{2}}}\right)\end{equation*}
\begin{equation*}2 = \exp\left({\lambda T_{\frac{1}{2}}}\right)\end{equation*}
\begin{equation*}\log{2} = \lambda T_{\frac{1}{2}}\end{equation*}
\begin{equation*}\lambda = \frac{log2}{T_{\frac{1}{2}}}\end{equation*}
We are told that the half-life is \(5\) days.
\begin{equation*}\lambda \approx 0.1386\end{equation*}
Substituting with the original mass gives us:
\begin{equation*}M_t = 800 e^{-0.1386t}\end{equation*}
where \(M_t\) is the mass in \(mg\) and \(t\) is the time in days.
b)
Find the mass remaining after \(30\) days.
Answer
\(N(30) = 12.5 mg\)
Solution
\begin{equation*}N(30) = 800 e^{-0.1386 \times 30} = 800 e^{-4.1589} = 800 \times 0.015625 = 12.5 \end{equation*}
c)
When is the mass reduced to \(1mg\)?
Answer
\(48.2\) days.
Solution
\begin{align*}
M_t&=M_0 e^{-\lambda t}\\
1&=800e^{-\lambda t}\\
\frac{1}{800}&=e^{-\lambda t}\\
800&=e^{\lambda t}\\
log{800}&=\lambda t\\
t&=\frac{log{800}}{\lambda}\\
&=\frac{log{800}}{0.1386}\\
& \approx 48.2
\end{align*}
d)
Sketch a graph of the mass function.
Answer
See Sage plot of mass function below.
Sketch of the mass function. (Note \(x,y\) are easier for plots)
If \(f''\) is continuous, show that:
\begin{equation*}\lim_{h \to 0}\frac{\fe{f}{x+h} - 2\fe{f}{x} + \fe{f}{x-h}}{h^2} = \fe{f''}{x}\end{equation*}
Definition3.7.1The Definition of Taylor Series
Given a smooth function \(f\), we can always write down a Taylor series;
there is no guarantee that the series converges to anything, let alone
to the function. Given a smooth function \(f : \mathbb{R} \to \mathbb{R}\), its Taylor
series (around \(0\)) is
\begin{equation*}\sum_{n=0}^\infty \frac{f^{(n)}(0) \, x^n}{n!}\end{equation*}
A common mistake is to use \(f^{(n)}(x)\) instead of \(f^{(n)}(0)\).
Given a smooth function \(f : \mathbb{R} \to \mathbb{R}\), its Taylor series expanded
around \(a\) is
\begin{equation*}\sum_{n=0}^\infty \frac{f^{(n)}(a) \, (x-a)^n}{n!}\end{equation*}
The first few entries are
\begin{equation*}f(a)+\frac {f'(a)}{1!} (x-a)+ \frac{f''(a)}{2!} (x-a)^2+\frac{f^{(3)}(a)}{3!}(x-a)^3+ \cdots\end{equation*}
a)
From first principles.
Solution
The first derivative is given by:
\begin{equation*}\frac{\Delta f(x)}{\Delta x} = \lim_{h \to 0}\frac{f(x+h)-f(x)}{(x+h)-(x)} = \lim_{h \to 0}\frac{f(x+h)-f(x)}{h}\end{equation*}
The solution is to apply this twice. The trick here is to start with the correct increments
(see Math Stack Q&A [3.7.3.1.1]).
\begin{align*}
\fe{f''}{x}&=\lim_{h \to 0} \frac{\fe{f'}{x} - \fe{f'}{x-h}}{h}\\
&=\lim_{h \to 0}\frac{\frac{f(x+h)-f(x)}{h}-\frac{f(x)-f(x-h)}{h}}{h}\\
&=\lim_{h \to 0}\frac{f(x+h)-2f(x)+f(x-h)}{h^2}
\end{align*}
b)
Using Taylor Series [3.7.3.1.2].
Solution
First set \(a=x+h\):
\begin{equation*}f(x)f(a)+\frac {f'(a)}{1!} (x-a)+ \frac{f''(a)}{2!} (x-a)^2+\frac{f^{(3)}(a)}{3!}(x-a)^3+ \cdots\end{equation*}
TBD
Subsubsection3.7.3.1References
[1]
Brian M. Scott
Second derivative “formula derivation”.
Math StackExchange
[2]
Suppose that \(3 \leq \fe{f'}{x} \leq 5\) for all values of \(x\), where
\(f\) is a function defined on all of the real numbers and differentiable everywhere.
Show that
\begin{equation*}18 \leq \fe{f}{8} - \fe{f}{2} \leq 30\end{equation*}a)
The Mean Value Theorem states that if \(f(x)\) is defined and continuous
on the interval \([a,b]\) and differentiable on \((a,b)\), then there is at least
one number \(c\) in the interval \((a,b)\) (that is \(a < c < b\)) such that
\begin{equation*}f'(c)=\frac{f(b) - f(a)}{b-a}\end{equation*}Solution
\begin{align*}
\text{Apply the mean value theorem}\\
&f'(c)=\frac{f(8) - f(2)}{8-2}\\
\text{Since we know}\\
&3 \leq \fe{f'}{x} \leq 5\\
\text{We get}\\
&3 \leq \frac{f(8) - f(2)}{6} \leq 5\\
&18 \leq f(8) - f(2) \leq 30
\end{align*}