🔬 Tutorial problems epsilon \(\epsilon\)#
Note
This problems are designed to help you practice the concepts covered in the lectures. Not all problems may be covered in the tutorial, those left out are for additional practice on your own.
\(\epsilon\).1#
Consider the function
(a) Does \(\lim _{x \rightarrow 5} f(x)\) exist? If so, what is it? Establish the validity of your answer formally using an epsilon-delta argument. If it exists, does it equal \(f(5)\) ? Is this function continuous at the point \(x=5\) ?
(b) Does \(\lim _{x \rightarrow 1} f(x)\) exist? If so, what is it? Establish the validity of your answer formally using an epsilon-delta argument. If it exists, does it equal \(f(1)\) ? Is this function continuous at the point \(x=1\) ?
(c) Is this function continuous on \(\mathbb{R}\)? Justify your answer.
Use examples in the lecture notes to establish \(\epsilon\)-\(\delta\) argument.
Consider the function
Part (a)
First of all, note that by the properties of the limits we have
Thus we conclude that (i) \(\lim _{x \rightarrow 5} f(x)=\frac{1}{4}\), and (ii), the function is continuous at the point \(x=5\). However, we also want to show that \(\lim _{x \rightarrow 5} f(x)=\frac{1}{4}\) formally, using an epsilon-delta proof.
Denote \(d(a, b)\) the Euclidean distance between \(a\) and \(b\) in \(\mathbb{R}\). We have
and
Suppose that we want \(d(f(5), f(x))<\varepsilon\), where \(\varepsilon>0\). Any problems will be caused by small values of \(\varepsilon>0\). (Try and explain why this is the case.) Thus we can focus on the cases where
Note that
This requires that both \(\frac{1}{(x-1)}-\frac{1}{4}<\varepsilon\) and \(-\left(\frac{1}{(x-1)}-\frac{1}{4}\right)<\varepsilon\).
We have
This requires that
Note that
We have
This requires that
Thus we know that
Suppose that we want \(d(5, x)<\delta\). This requires that
Thus we want to choose \(\delta\) in such a way that
This can always be done, so we have
Part (b)
It can be shown that \(\lim _{x \rightarrow 1} f(x)\) does not exist. This means that \(f(x)\) is discontinuous at the point \(x=1\).
Suppose that \(\lim _{x \rightarrow 1} f(x)=l\) for some \(l \in \mathbb{R}\). We will consider three distinct cases. The first case will involve \(l=0\). The second case will involve \(l>0\). The third case will involve \(l<0\).
Case 1: \(l=0\).
Suppose that \(\lim _{x \rightarrow 1} f(x)=0\). In order for \(\lim _{x \rightarrow 1} f(x)=0\), we need there to exist a function of \(\varepsilon, \delta: \mathbb{R}_{++} \longrightarrow \mathbb{R}_{++}\), such that, for each \(\varepsilon>0\), we have \(d(f(x), 0)<\varepsilon\) whenever \(d(x, 1)<\delta(\varepsilon)\). If we can find so much as a single \(\varepsilon>0\) for which no such \(\delta(\varepsilon)\) exists, then the limit does not exist. Note that
Let \(\varepsilon>0\). If \(x>1\), we have
because \(x-1>0\) when \(x>1\). On the other hand, if \(x<1\), we have
because \(1-x>0\) when \(x<1\). Recall that \(d(x, 1)=|x-1|\). Thus, when \(x \neq 1\) we know that \(d(f(x), 0)<\varepsilon\) if and only if \(d(x, 1)>\frac{1}{\varepsilon}\). Thus we know that, for any \(\varepsilon>0\), there is no \(\delta(\varepsilon)>0\) such that \(d(f(x), 0)<\varepsilon\) for all \(d(x, 1)<\delta(\varepsilon)\). This means that \(\lim _{x \rightarrow 1} f(x) \neq 0\). (This alone means that the function \(f(x)=\frac{1}{x-1}\) is not continuous at the point \(x=1\), because \(f(1)=0\).)
Case 2: \(l>0\).
Suppose that \(\lim _{x \rightarrow 1} f(x)=l>0\). In order for \(\lim _{x \rightarrow 1} f(x)=l>0\), we need there to exist a function of \(\varepsilon, \delta: \mathbb{R}_{++} \longrightarrow \mathbb{R}_{++}\), such that, for each \(\varepsilon>0\), we have \(d(f(x), l)<\varepsilon\) whenever \(d(x, 1)<\delta(\varepsilon)\). If we can find so much as a single \(\varepsilon>0\) for which no such \(\delta(\varepsilon)\) exists, then the limit does not exist. Note that
What happens when \(x=1\) here is irrelevant. What happens when \(x\) gets arbitrarily close to one is very relevant.
Let \(\varepsilon>0\). Suppose that \(f(x)>l>0\). Note that if \(f(x)>0\), we must have \(x>1\). In such a situation, we have
because (i) \(x-1>0\) when \(x>1\), and (ii) \(l+\varepsilon>0\) when both \(\varepsilon>0\) and \(l>0\).
Let \(\varepsilon>0\). In fact, we will restrict attention to the subset of cases in which \(0<\varepsilon<l\). Suppose that \(0<f(x)<l\). Note that if \(f(x)>0\), we must have \(x>1\). In such a situation, we have
because (i) \(x-1>0\) when \(x>1\), and (ii) \(l-\varepsilon>0\) when \(0<\varepsilon<l\).
Let \(\varepsilon>0\). In fact, we will restrict attention to the subset of cases in which \(0<\varepsilon<l\). Note that if \(x<1\), we must have \(f(x)<0<l\). This means that, in cases where \(x<1\), we have
because (i) \(x-1<0\) when \(x<1\), and (ii) \(l-\varepsilon>0\) when \(0<\varepsilon<l\).
We have thus established that when \(l>0\) and \(\varepsilon \in(0, l)\), we will have \(d(f(x), l)<\varepsilon\) if and only if \(d(x, 1)>\max \left(\frac{1}{l+\varepsilon}, \frac{1}{l-\varepsilon}\right)=\frac{1}{l-\varepsilon}>0\). Thus we know that, for any \(\varepsilon \in(0, l)\), there is no \(\delta(\varepsilon)>0\) such that \(d(f(x), l)<\varepsilon\) for all \(d(x, 1)<\delta(\varepsilon)\). This means that \(\lim _{x \rightarrow 1} f(x) \neq l\) for any \(l>0\).
Case 3: \(l<0\).
This case is left as an exercise. An appropriate variation on the arguments used for case 2 should work.
Combining the results from each of the three cases, we can conclude that the limit of the function as \(x\) approaches one does not exist. Since the limit of the function as \(x\) approaches one does not exist, we know that the function must be discontinuous at the point \(x=1\).
Part (c)
Since \(f(x)\) is discontinuous at the point \(x=1\), we know that it is a discontinuous function on \(\mathbb{R}\).
\(\epsilon\).2#
Consider the function
(a) Does \(\lim _{x \rightarrow 0} f(x)\) exist? If so, what is it? Try and establish the validity of your answer formally using an epsilon-delta argument. If it exists, does it equal \(f(0)\) ? Is this function continuous at the point \(x=0\) ?
(b) Is this function continuous?
Consider the function
Part (a)
It can be shown that \(\lim _{x \rightarrow 0} f(x)\) exists, but is equal to zero, so that
This means that \(f(x)\) is discontinuous at the point \(x=0\).
Denote \(d(a, b)\) the Euclidean distance between \(a\) and \(b\) in \(\mathbb{R}\). We have
and
We want to find some \(\delta_{\varepsilon}>0\) for some arbitrary \(\varepsilon>0\), such that, whenever \(d(0, x)=|x|<\delta_{\varepsilon}\), we have \(d(0, f(x))=|f(x)|<\varepsilon\). We are interested in the behaviour of \(f(x)\) as \(x\) approaches zero here, not in the actual value taken by \(f(x)\) at the point \(x=0\). As such, we have
Consider some \(\varepsilon>0\). We want \(d(0, f(x))<\varepsilon\). Note that
Thus if we set \(\delta_{\varepsilon}=\sqrt{\varepsilon}\), then we know that \(d(0, f(x))<\varepsilon\) whenever \(d(0, x)<\delta_{\varepsilon}\). This means that \(\lim _{x \rightarrow 0} f(x)=0\).
Note that \(\lim _{x \rightarrow 0} f(x)=0 \neq-50=f(0)\). This means that the function \(f(x)\) is discontinuous at the point \(x=0\).
Part (b)
Since the point \(x=0\) belongs to the domain of the function \(f(x)\) and \(f(x)\) is discontinuous at the point \(x=0\), we know that it is a discontinuous function.
\(\epsilon\).3#
Consider the function
(a) Is this function continuous at the point \(x=2\) ? Justify your answer.
(b) Is this function continuous? Justify your answer.
Consider the function
Part (a)
We will show that
so that the function \(f(x)\) is continuous at the point \(x=2\).
Denote \(d(a, b)\) the Euclidean distance between \(a\) and \(b\) in \(\mathbb{R}\). We have
and
Let \(\varepsilon>0\) and \(x<2\). Note that
This requires that either (i) \(x^{2}-4<\varepsilon\), so that \(x^{2}<4+\varepsilon\), or (ii) \(-\left(x^{2}-4\right) < \varepsilon\), so that \(x^{2}>4-\varepsilon\). Thus we have
This means that, if \(\varepsilon>0\) and \(x<2\), then \(d(0, f(x))<\varepsilon\) requires that either
or
Since we require that \(x>2\) in this case, we actually have
(i) either \(\sqrt{4-\varepsilon}<x<2\) or (ii) \(-\sqrt{4+\varepsilon}<x<-\sqrt{4-\varepsilon}\).
Let \(\varepsilon>0\) and \(x \geqslant 2\). Note that
Since we require that \(x \geqslant 2\) in this case, we actually have
Upon combining our results, we can conclude that \(d(0, f(x))<\varepsilon\) if \(\sqrt{4-\varepsilon}<\) \(x<2\). Let
and let \(\delta_{\varepsilon}=\min \left(\varepsilon, \gamma_{\varepsilon}\right)\). Note that \(\gamma_{\varepsilon}>0\), so that \(\delta_{\varepsilon}>0\). This ensures that \(d(0, f(x))<\varepsilon\) whenever \(d(2, x)<\delta_{\varepsilon}\). As such, we know that
so that the function \(f(x)\) is continuous at the point \(x=2\).
Part (b)
We have shown that the function
is continuous at the point \(x=2\). In order to show that the function is continuous, we also need to show that the function \(f(x)=x^{2}+1\) is continuous for all \(x<2\), and that the function \(f(x)=x+3\) is continuous for all \(x>2\). This can indeed be done, so that the function is indeed continuous. However, these additional steps are left as an exercise.
\(\epsilon\).4#
Use either the product rule or the quotient rule to find the first derivative of each of the following functions.
(a) \(y=x^{3}\left(x^{2}+4\right)\);
(b) \(y=(x-3)\left(x^{2}-5 x+7\right)\);
(c) \(y=\frac{\left(x^{3}+6 x^{2}-2\right)}{x^{4}}\);
(d) \(y=\left(x^{4}+3 x\right) x^{-6}\);
(e) \(y=5+x e^{x}\);
(f) \(y=x^{2} e^{x}+x \ln (x)\);
(g) \(y=x^{2}(1.1)^{x}\);
(h) \(y=5 e^{x} \ln (x)\);
(i) \(y=\left(x^{3}+2\right)(1.2)^{x}\); and
(j) \(y=x^{-\left(\frac{5}{6}\right)} e^{x}\).
[Shannon, 1995], p. 401
(a)
(b)
(c)
(d)
(e)
(f) In this example the product rule is used to find the derivative of both terms on the RHS
(g)
(h)
(i)
(j)
\(\epsilon\).5#
Use the chain rule to find the first derivative of each of the following functions.
(a) \(y=(3 x+4)^{3}\);
(b) \(y=\left(2 x^{2}+6\right)^{4}\);
(c) \(y=(3-2 x)^{-2}\);
(d) \(y=\left(4+\ln \left(x^{2}\right)\right)^{-1}\);
(e) \(y=5+e^{x^{2}}\);
(f) \(y=10 x+e^{\ln (x)}\);
(g) \(y=\ln \left(x+x^{\frac{3}{2}}\right)\);
(h) \(y=\frac{10}{\left(1-5 e^{0.2 x}\right)}\);
(i) \(y=\sqrt{\frac{8}{x^{-1}}}\); and
(j) \(y=100-5 e^{-0.3 x}\).
[Shannon, 1995], p. 401
The chain rule or function of a function rule is
(a) \(y=(3 x+4)^{3}=u^{3}\) where \(u=3 x+4\)
(b) \(y=\left(2 x^{2}+6\right)^{4}=u^{4}\) where \(u=2 x^{2}+6\)
(c) \(y=(3-2 x)^{-2}=u^{-2}\) where \(u=3-2 x\)
(d) \(y=\left(4+\operatorname{ln} x^{2}\right)^{-1}=u^{-1}\) where \(u=4+\operatorname{ln} x^{2}\)
(The chain rule was used to find the derivative of \(\operatorname{ln} x^{2}\).)
(e) As the derivative of 5 is 0 we only need use the chain rule to find the derivative of \(\mathrm{e}^{\mathrm{x}^{2}}\)
\(y =e^{x^{2}}=e^{u(x)} \text { where } u=x^{2} \\\)
(f) The derivative of \(10x\) is \(10(1)=10\) so the chain rule is only needed to find the derivative of \(\mathrm{e}^{\ln x}\)
\(y =e^{\ln x}=e^{u(x)} \text { where } u=\ln x \\\)
The derivative of the original function is \(10+1\) or \(11\).
(g) \(\quad y=\operatorname{ln}\left(x+x^{3 / 2}\right)=\operatorname{ln} u\) where \(\mathrm{u}=x+x^{3 / 2}\)
(h) \(\quad y=\frac{10}{\left(1-e^{2 x}\right)}=10 u^{-1}\) where \(u=1-5 e^{2 x}\)
(i) \(y=\sqrt{8 x^{-1}}=u^{1 / 2}\), where \(u=8 x^{-1}\)
assuming that \(x \neq 0\).
(j) The derivative of 100 is 0 so we only use the chain rule to find the derivative of \(-5 e^{-0.3 x}\)
\(y=-5 e^{-0.3 x}=-5 e^{u(x)} \text { where } u=-0.3 x\)
\(\epsilon\).6#
Consider a function \(f: X \longrightarrow \mathbb{R}\) where \(X \subseteq \mathbb{R}\). Suppose that \(\lim _{h \rightarrow 0}\left(\frac{f(a+h)-f(a)}{h}\right)\) exists. Show that this implies that \(f(x)\) is continuous at the point \(a \in X\).
Consider a function \(f: X \longrightarrow \mathbb{R}\) where \(X \subseteq \mathbb{R}\). Suppose that
exists. We want to show that this implies that \(f(x)\) is continuous at the point \(a \in X\).
This proof of this proposition is drawn from Ayers and Mendelson (2013, Chapter 8, Solved Problem 2).
First, note that
Thus we have
Now note that
Upon combining these two results, we obtain
Finally, note that
Thus we have
This means that \(f(x)\) is continuous at the point \(x=a\).
\(\epsilon\).7#
Approximate the function
around \(x = 1\) using the Taylor series.
Provide:
The linear approximation
The quadratic approximation
The cubic approximation
Use each approximation to estimate the value of \(f(0)\). Is this an accurate estimate?
In other words, use as many terms of Tayor series to build an approximation by a linear function, quadratic and cubic polynomials.
We use the Taylor expansion formula about \(x = 1\):
#Step 1: Evaluate function and derivatives at \(x = 1\)
First derivative:
\[ f'(x) = -\frac{1}{x^2} e^{1/x} + \frac{1}{x} \quad \Rightarrow \quad f'(1) = -e + 1 = 1 - e \]Second derivative:
\[ f''(x) = \left( \frac{2}{x^3} + \frac{1}{x^4} \right) e^{1/x} - \frac{1}{x^2} \quad \Rightarrow \quad f''(1) = 3e - 1 \]Third derivative:
\[ f^{(3)}(x) = \left( -\frac{6}{x^4} - \frac{6}{x^5} - \frac{1}{x^6} \right) e^{1/x} + \frac{2}{x^3} \quad \Rightarrow \quad f^{(3)}(1) = -13e + 2 \]
Linear Approximation
Using up to the first derivative:
Substitute values:
Compute approximation:
Quadratic Approximation
Include up to the second derivative:
Compute approximation using previous approximation:
Cubic Approximation
Include up to the third derivative:
Compute approximation using previous approximation:
All approximations are rather inaccurate because the function is undefined at \(x = 0\).
But as \(\lim_{x \to 0^+} f(x) = \infty\), so the approximated values are increasing with the order of the approximation.