202008181129 Homework 1 (Q1)

Solve for z\in \mathbb{C} in the following equations:

(a) z^4+z^3+z^2+z+1=0,

(b) 3z^3+29z^2+497z-169=0.


 Attempts.

(a) Take notice that z\neq 1. Try and see having both sides of the equation multiplied by (1-z),

\begin{aligned} 0 & = (1-z)(z^4+z^3+z^2+z+1)  \\ 0 & = 1-z^5  \\ z^5 & = 1 = 1(1+0\,\mathrm{i}) \\ z^5 & = e^{\mathrm{i}(2n\pi)}\qquad \qquad \textrm{where }n=0,1,2,3,4. \\ \end{aligned}

\therefore z=e^{\mathrm{i}(\frac{2n\pi}{5})}.

(n=0 is rejected for z\neq 1=e^{\mathrm{i}(\frac{2(0)\pi}{5})}.)

In polar expression z=e^{\mathrm{i\frac{2\pi}{5}}},\, e^{\mathrm{i\frac{4\pi}{5}}},\, e^{\mathrm{i\frac{6\pi}{5}}},\, e^{\mathrm{i\frac{8\pi}{5}}}.

In trigonometric expression

z=\mathrm{cis}(\frac{2n\pi}{5})=\mathrm{cis}(\frac{2\pi}{5}),\, \mathrm{cis}(\frac{4\pi}{5}),\, \mathrm{cis}(\frac{6\pi}{5}),\, \mathrm{cis}(\frac{8\pi}{5})

I.e.,

\begin{aligned} z^1 & = \cos 72^\circ +\mathrm{i}\sin 72^\circ = \mathrm{cis\,} 72^\circ \\ z^2 & = \cos 144^\circ +\mathrm{i}\sin 144^\circ = \mathrm{cis\,} 144^\circ \\ z^3 & = \cos 216^\circ +\mathrm{i}\sin 216^\circ = \mathrm{cis\,} 216^\circ \\ z^4 & = \cos 288^\circ +\mathrm{i}\sin 288^\circ = \mathrm{cis\,} 288^\circ \\ \end{aligned}

(b)Let f(z)=3z^3+29z^2+497z-169=0. Then (3z-1) is a factor, because \frac{1}{3} is a zero (i.e., f(\frac{1}{3}) = \frac{1}{9}+\frac{29}{9}+\frac{497}{3}-169=0).

\begin{aligned} f(z) & =(3z-1)(z^2+10z+169) \\ & = (3z-1)g(z) \\ \end{aligned}

When g(z)=0, z=\displaystyle{\frac{-10\pm \sqrt{100-4(169)}}{2}}=-5\pm 12\,\mathrm{i}.

The solution to f(z)=0 gives z_1=\frac{1}{3}, z_2=-5+12\,\mathrm{i}, and z_3=-5-12\,\mathrm{i}.

202004241907 Homework 1 (Q2)

(a) State the Mean Value Theorem and the Taylor’s Theorem.

(b) Find an O(h^4) and an O(h^5) approximation to \cos h. Compare those approximate values to the actual value 0.9950042 when h=0.1. Correct answers in this question to seven decimal places.


Solution.

(a)

Theorem. (Mean-Value Theorem) Suppose f\in C[a,b], and f'(x) exists on (a,b). For every x\in [a,b] and some x_0\in [a,b],

f(x)=f(x_0)+f'(\eta (x))(x-x_0)

where \eta (x) is between x_0 and x.

Theorem. (Taylor’s Theorem) Suppose f\in C^n[a,b], and f^{(n+1)} exists on (a,b). For every x\in (a,b) and some x_0\in [a,b], there exists a number \eta (x) between x_0 and x with f(x)=P_n(x)+R_n(x), where

P_n(x)=f(x_0)+f'(x_0)(x-x_0)+ \displaystyle{\frac{f''(x_0)}{2!}}(x-x_0)^2+\cdots + \displaystyle{\frac{f^{(n)}(x_0)}{n!}(x-x_0)^n}, and

R_n(x)=\displaystyle{\frac{f^{(n+1)}(\eta (x))}{(n+1)!}}(x-x_0)^{n+1}.

(b) Let f(x)=\cos x. Then we have f'(x)=-\sin x, f''(x)=-\cos x, f'''(x)=\sin x, f^{(4)}(x)=\cos x, f^{(5)}(x)=-\sin x, etc. By Taylor expansion of \cos x at x_0, we have

\cos x=\cos x_0+(-\sin x_0)(x-x_0)+\displaystyle{\frac{-\cos x_0}{2!}}(x-x_0)^2+\displaystyle{\frac{\sin x_0}{3!}}(x-x_0)^3+\displaystyle{\frac{\cos x_0}{4!}}(x-x_0)^4+\displaystyle{\frac{-\sin x_0}{5!}}(x-x_0)^5+\cdots

Now letting x_0=0, we obtain further that

\cos x=1-\displaystyle{\frac{x^2}{2!}}+\displaystyle{\frac{x^4}{4!}}+\cdots =\displaystyle{\sum_{n=0}^\infty \frac{(-1)^n}{(2n)!}x^{2n}}

Thus, the O(h^4) approximation to \cos h is in the form of

\cos h=1-\displaystyle{\frac{h^2}{2!}}+O(h^4)

where O(h^4) is the higher-order terms \displaystyle{\sum_{n=2}^\infty \frac{(-1)^n}{(2n)!}h^{2n}}.

And the O(h^5) approximation to \cos h is in the form of

\cos h=1-\displaystyle{\frac{h^2}{2!}}+\displaystyle{\frac{h^4}{4!}}+O(h^5)

where O(h^5) is the higher-order terms \displaystyle{\sum_{n=3}^\infty \frac{(-1)^n}{(2n)!}h^{2n}}.

Then, when h=0.1, \cos h in the O(h^4) approximation would be

\cos 0.1=1-\displaystyle{\frac{0.1^2}{2!}}=0.995;

whereas in the O(h^5) approximation,

\cos 0.1=1-\displaystyle{\frac{0.1^2}{2!}}+\displaystyle{\frac{0.1^4}{4!}}=0.995004166\approx 0.9950042\enspace \textrm{(corr. to 7 d.p.)}.

Given the actual value \cos h=0.9950042, the O(h^4) approximation has an absolute error of |0.9950042-0.995|=0.0000042 and a relative error of \displaystyle{\frac{|0.9950042-0.995|}{|0.9950042|}}=4.2210877\times 10^{-6}. And the O(h^5) approximation has an absolute error of |0.9950042-0.9950042|=0 and thus a relative error of zero.


Remark. The 0 error is owing to the correction of 7 decimal places in both the direct computation of \cos 0.1 and its O(h^5) approximation.

202004241649 Problem 2, Ch. 1 Sec. 1

Prove the following equalities:

(a) |ab|=|a|\cdot |b|;

(b) |a|^2=a^2;

(c) \displaystyle{\bigg|\frac{a}{b}\bigg|=\frac{|a|}{|b|}} (where b\neq 0);

(d) \sqrt{a^2}=|a|.


Proof.

(a) (proof by cases)

i. When a,b\geqslant 0:

|ab|=ab=|a||b|.

ii. When a,b< 0:

|ab|=ab=(-a)(-b)=|a||b|.

iii. When a\geqslant 0 and b<0:

|ab|=-ab=a(-b)=|a||b|.

iv. When a<0 and b\geqslant 0:

|ab|=-ab=(-a)b=|a||b|.

QED


(b) (proof by induction)

By making a stronger claim

P(n): For any positive integer n, |a^n|=|a|^n.

Proof.

The trivial cases n=0 and n=1 are evident.

Consider the case n=2,

\begin{aligned} |a^2| & = |a||a| \qquad \textrm{(by equality (a))} \\ & = |a|^2 \end{aligned}

P(2) is true.

Assume now that P(n) is true,

\begin{aligned} P(n+1): \qquad |a^{n+1}|& = |a^n\cdot a| \\ & = |a^n||a| \qquad \textrm{(by equality (a))}\\ & = |a|^n|a| \qquad \textrm{(by the assumption }P(n)\textrm{ is true)} \\ & = |a|^{n+1} \end{aligned}

it can be seen that P(n+1) is also true.

From the fact that P(2) is true and by the principle of mathematical induction, P(n) is true for all positive integers n.

It follows that |a|^2=|a^2|=a^2 holds.

QED


(c) (direct proof)

\because |a|=\displaystyle{\bigg|  \frac{a}{b}\cdot b\bigg|} =\bigg|\displaystyle{\frac{a}{b}}\bigg| |b|,

where the second equality sign is due to equality (a),

\therefore \displaystyle{\bigg| \frac{a}{b} \bigg| = \frac{|a|}{|b|}}.

QED


(d) (proof by definition)

The absolute value of a real number a, denoted by |a|, is defined by

|a|= \begin{cases} x & \textrm{if } x\geqslant 0, \\  -x & \textrm{if }x<0  \end{cases}

For any non-negative real number a, the symbol \sqrt{a} denotes the non-negative square root of a.

QED

202004241533 Problem 1, Ch. 1 Sec. 1

Prove that if a and b are real numbers then

||a|-|b||\leqslant |a-b|\leqslant |a|+|b|.


Proof.

As |a| is non-negative and -|a| non-positive, one has

Eq. (1):

-|a|\leqslant a\leqslant |a|

Eq. (2):

-|b|\leqslant b\leqslant |b|

Combining Eq. (1) and Eq. (2),

-(|a|+|b|)\leqslant a+b\leqslant |a|+|b|,

or, Eq. (3): (the triangle inequality)

|a+b|\leqslant |a|+|b|.

Applying the triangle inequality to |a-b|, one gets

Eq. (4):

|a-b|=|a+(-b)|\leqslant |a|+|-b|=|a|+|b|,

or, Eq. (4)’:

|a-b|\leqslant |a|+|b|.

Applying the triangle inequality to |a|=|(a-b)+b|, one gets

Eq. (5):

|a|=|(a-b)+b|\leqslant |a-b|+|b|,

or, Eq. (5)’:

|a|-|b|\leqslant |a-b|.

Applying the triangle inequality to |b|=|(b-a)+a|, one gets

Eq. (6):

|b|=|(b-a)+a|\leqslant |b-a|+|a|,


Roughwork.

\begin{aligned} |b|-|a| & \leqslant |b-a| \\|b|-|a| & \leqslant |a-b| \\-|a-b| & \leqslant |a|-|b|\end{aligned}


or, Eq. (6)’:

-|a-b| \leqslant |a|-|b|

Combining Eq. (5)’ and Eq. (6)’:

-|a-b| \leqslant |a|-|b| \leqslant |a-b|,

or, Eq. (7):

||a|-|b||\leqslant |a-b|.

Combining Eq. (4)’ and Eq. (7), one obtains readily

||a|-|b||\leqslant |a-b|\leqslant |a|+|b|.

QED

202004240713 Homework 1 (Q3)

Suppose G be a finite group of even order. Prove that there exists an element a\in G such that a\neq e and a^2=e.


Attempts.

Try negating the statement by the following:

\forall\, a\in G, either a=e or a^2\neq e,

then looking for contradiction to the assumption that G should be a finite group of even order.


CASE 1: Would it be possible that a=e?

Given so, the inverse and the identity of G would be e. And the set G would have contained e only, i.e., the singleton set G=\{ e\}. This contradicts with the assumption that its order be even.

CASE 2: If for all elements in the group G, their order cannot ever be 2.

If a^2\neq e, then a\neq a^{-1} for all a\in G. It would then become an ill-posited negation, because the identity element e is one of all a\in G, and both statements e^2\neq e and e\neq e^{-1} are abhorrent to the basic axioms of a group. So perhaps to my discretion e should be precluded from a‘s, and understood as one (e=e^{-1}) member apart from the other members.

Proceeding to observe that if a\neq a^{-1} for all a\in G, I may then construct a pair of a_i and a_i^{-1} for i\in \{ 1,2,\dots ,n:n\in\mathbb{Z}^+\}, provided from the definition of a group that the inverse of each element must exist and, as it follows, that it must be unique. That is, if a_5=a_{27}^{-1}, I may relabel a_{27} as a_5^{-1}. By such a construction I can guarantee that all elements, except the identity e, are now in pairs.

One therefore, by counting in total how many elements there are in group G, will get an odd number (2n+1), the 2n counted from the pairs, and the single 1 the identity e itself alone counts.

This contradicts with the assumption that |G| be of even parity.

From the negation of statement arises contradiction, the original statement is therefore proven by contradiction.

Remark.

I could not prove otherwise than loosely to so naive myself. Not until I had survived mathematical rigor.

202004240630 Homework 1 (Q2)

Call G a group and e the identity of G. If the order of G is 2, then a^2=e for all a\in G and G is abelian (Prove it yourself). Show that

G is abelian if a^2=e for all a\in G.

Give an example of such a group whose order is greater than 2.


Solution.

Recall that a group G is said to be abelian if xy=yx for all x,y\in G.

For any a,b\in G, there is ab=(bb^{-1})ab(a^{-1}a), because a group satisfies, first, the inverse axiom xx^{-1}=x^{-1}x=e under the notation x^{-1} being the inverse of x, and secondly, the identity axiom ex=xe=x.

Thence by (associativity axiom) x*(y*z)=(x*y)*z I may change brackets and obtain as follows:

\begin{aligned} ab & =(bb^{-1})ab(a^{-1}a) \\ & =b(b^{-1}a)(ba^{-1})a \\ & =b(b^{-1}a)(b^{-1}a)^{-1}a \end{aligned}

Provided that a^2=e for all a\in G, one may deduce a=a^{-1}, b=b^{-1}, etc. Thus,

ab=b(b^{-1}a)(b^{-1}a)a

In addition (b^{-1}a)(b^{-1}a)=(b^{-1}a)^2=e, hence ab=ba, and G abelian.

The product group \mathbb{Z}_2\times \mathbb{Z}_2 is an example of one abelian group whose order is greater than 2 and the order of whose elements is 2.

202004240606 Homework 1 (Q1)

Let G=\{ a,b,c,d\} and the Cayley table of G be

\begin{aligned} * \quad | &\textrm{} \quad a &\textrm{}\quad b \quad &\textrm{}\quad c &\textrm{}\quad d \\ \hline a \quad | &\quad a &\quad b \quad &\quad c &\quad d \\ b \quad | &\quad b &\quad a \quad &\quad c &\quad d \\ c \quad | &\quad c &\quad b \quad &\quad a &\quad d \\ d \quad | &\quad d &\quad d \quad &\quad b &\quad c \end{aligned}

Is (G,*) a group?


Attempts.

I recall that a group (G,*) is a set G altogether with an operation *:G\times G\rightarrow G, (x,y)\mapsto x*y such that the following axioms be satisfied:

i. (Associativity) (x*y)*z=x*(y*z);

ii. (Identity) There exists e\in G such that e*x=x*e=x for all x\in G; and

iii. (Inverse) For each a\in G, there exists b\in G such that a*b=b*a=e.


Let me check them one by one.

i. For x,y,z\in \{ a,b,c,d \}, it suffices to check 4\times 4\times 4=64 operations. And I found that the following operation failed axiom (i):

d*(c*b)=d*b=d but (d*c)*b=b*b=a.

That said, G under the operation * is not a group.

ii. By inspection, the identity of G exists, and that is a because a*x=x*a=x for all x\in \{ a,b,c,d\}. Axiom (ii) is thus satisfied.

iii. There exists d\in G, such that for all x^i\in\{ a,b,c,d\}=G,

d*x^i=x^i*d\neq a.

The claim above is validated by simply checking the identity found to be a in part ii. cannot be found in the entries, and by the fact that a is unique.


Axioms i. and iii. having been failed, I may conclude that G under the operation * cannot form a group.

202004231606 Exercise 1, Section 1.1

Determine whether the vectors emanating from the origin and terminating at the following pair of points are parallel.

(a) (3,1,2) and (6,4,2)

(b) (-3,1,7) and (9,-3,-21)

(c) (5,-6,7) and (-5,6,-7)

(d) (2,0,-5) and (5,0,-2)


Background.

Two nonzero vectors x and y are called parallel if y=tx for some nonzero real number t. (Thus nonzero vectors having the same or opposite directions are parallel.)

Text on pg.3


Solution.

(a) Let x=(3,1,2) and y=(6,4,2). Apparently \nexists\, t\in \mathbb{R} such that y=tx. For otherwise (6,4,2)=t(3,1,2), the system of equations

\begin{aligned} 6 & = 3t \\ 4 & = 1t \\ 2 & = 2t \\ \end{aligned}

is inconsistent. They are not parallel.

(b) Let x=(-3,1,7) and y=(9,-3,-21), then y=-3x. The vectors x and y are in opposite direction and the magnitude of y is three times that of x. They are parallel.

(c) Let x=(5,-6,7) and y=(-5,6,-7). Observe that they are in equal magnitude but in opposite direction, i.e., y=-x. They are also parallel.

(d) Let x=(2,0,-5) and y=(5,0,-2). Assume t\in\mathbb{R} s.t. y=tx, i.e.,

\begin{aligned} 5 & = 2t \\ 0 & = 0t \\ -2 & = 5t \\ \end{aligned}

no way will the first and the third lines agree. They are nonparallel.

201911190336 Homework 1, Differential Geometry

Let \boldsymbol{\alpha} :I\rightarrow \mathbb{R}^3 be a regular parametrized curve (not necessarily by arc length) and let \beta :J\rightarrow \mathbb{R}^3 be a reparametrization of \boldsymbol{\alpha} by the arc length s=s(t) measured from t_0\in I.

Let also t=t(s) be the inverse function of s and denote the derivative of \boldsymbol{\alpha} wrt t by \boldsymbol{\alpha}'. Prove that

i. \mathrm{d}t/ \mathrm{d}s=1/\|\boldsymbol{\alpha}'\| and \mathrm{d}^2t/ \mathrm{d}s^2=-\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}''\rangle /\|\boldsymbol{\alpha}'\|^4;

ii. The curvature of \boldsymbol{\alpha} at t\in I is \kappa (t)=\displaystyle{\frac{\|\boldsymbol{\alpha}' \wedge\boldsymbol{\alpha}''\|}{\|\boldsymbol{\alpha}'\|^3}}; and

iii. The torsion of \boldsymbol{\alpha} at t\in I is \tau (t)=\displaystyle{-\frac{\langle \boldsymbol{\alpha}' \wedge \boldsymbol{\alpha}'',\boldsymbol{\alpha}'''\rangle }{\| \boldsymbol{\alpha}' \wedge\boldsymbol{\alpha}''\|^2}}.


Solution.

i. By definition \displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}t}}=\boldsymbol{\alpha}'.

Using chain rule,

\displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}t}}= \displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}s}}\displaystyle{\frac{\mathrm{d}s}{\mathrm{d}t}}.

After taking the norm, as a consequence of natural parametrization

(i.e., \bigg| \displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}s}}\bigg| =1),

we have

\| \boldsymbol{\alpha}'\|=\bigg| \displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}s}} \bigg| \bigg| \displaystyle{\frac{\mathrm{d}s}{\mathrm{d}t}} \bigg|= \bigg| \displaystyle{\frac{\mathrm{d}s}{\mathrm{d}t}}\bigg|.

Hence \mathrm{d}t/\mathrm{d}s=1/ \| \boldsymbol{\alpha}'\|.

That said,

\begin{aligned} \displaystyle{\frac{\mathrm{d}^2t}{\mathrm{d}s^2}} & =\bigg( \displaystyle{\frac{\mathrm{d}t}{\mathrm{d}s}} \bigg) \displaystyle{\frac{\mathrm{d}}{\mathrm{d}t}}\bigg( \displaystyle{\frac{\mathrm{d}t}{\mathrm{d}s}} \bigg) \\ & =\displaystyle{\frac{1}{\| \boldsymbol{\alpha}' \|}}\frac{\mathrm{d}}{\mathrm{d}t}\bigg(\displaystyle{\frac{1}{\|\boldsymbol{\alpha}' \|}}\bigg) \\ & =\displaystyle{\frac{1}{\| \boldsymbol{\alpha}' \|}}\frac{\mathrm{d}}{\mathrm{d}t}\bigg(\displaystyle{\frac{1}{\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{1/2}}}\bigg)\\ \end{aligned}.

But,

\begin{aligned} & \displaystyle{\frac{\mathrm{d}}{\mathrm{d}t}}\bigg(\displaystyle{\frac{1}{\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{1/2}}}\bigg)\\ & =-\displaystyle{\frac{1}{2}}\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{-3/2}\Big( \langle \boldsymbol{\alpha}',\boldsymbol{\alpha}'' \rangle +\langle \boldsymbol{\alpha}'',\boldsymbol{\alpha}' \rangle \Big) \\ & =-\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{-3/2}\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}'' \rangle \\ \end{aligned}.

Thus,

\begin{aligned} \displaystyle{\frac{\mathrm{d}^2t}{\mathrm{d}s^2}} & =-\displaystyle{\frac{1}{\| \boldsymbol{\alpha}' \|}}\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{-3/2}\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}'' \rangle \\ & =-\displaystyle{\frac{1}{\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{1/2}}}\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}' \rangle^{-3/2}\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}'' \rangle \\ & = \displaystyle{-\frac{\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}'' \rangle}{\big( \langle \boldsymbol{\alpha}', \boldsymbol{\alpha}'\rangle^{1/2}\big)^4}} \\ & =-\langle \boldsymbol{\alpha}',\boldsymbol{\alpha}''\rangle /\|\boldsymbol{\alpha}'\|^4 \\ \end{aligned}.

ii.

\boldsymbol{\alpha}'= \displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}t}}=\displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}s}}\displaystyle{\frac{\mathrm{d}s}{\mathrm{d}t}}=\dot{\boldsymbol{\alpha}}s'.

Besides,

\boldsymbol{\alpha}''=\displaystyle{\frac{\mathrm{d}}{\mathrm{d}t}}(\dot{\boldsymbol{\alpha}}s')=\dot{\boldsymbol{\alpha}}\displaystyle{\frac{\mathrm{d}s'}{\mathrm{d}t}}+s'\displaystyle{\frac{\mathrm{d}\dot{\boldsymbol{\alpha}}}{\mathrm{d}t}}=\dot{\boldsymbol{\alpha}}s''+(s')^2\ddot{\boldsymbol{\alpha}}.

Then,

\langle \boldsymbol{\alpha}'\wedge \boldsymbol{\alpha}'' \rangle =\big( \dot{\boldsymbol{\alpha}}s' \big) \times \big( \dot{\boldsymbol{\alpha}}s''+(s')^2\ddot{\boldsymbol{\alpha}} \big) =(s')^3\langle\dot{\boldsymbol{\alpha}}\wedge \ddot{\boldsymbol{\alpha}} \rangle =\|\boldsymbol{\alpha}'\|^3 \langle \dot{\boldsymbol{\alpha}}\wedge \ddot{\boldsymbol{\alpha}} \rangle.

(For \mathrm{d}s/\mathrm{d}t=\| \boldsymbol{\alpha}'\| of part i. is used.)

It follows that

\langle \boldsymbol{\alpha}'\wedge \boldsymbol{\alpha}'' \rangle =\|\boldsymbol{\alpha}\|^3 \| \dot{\boldsymbol{\alpha}} \|\| \ddot{\boldsymbol{\alpha}}\|\sin \measuredangle (\dot{\boldsymbol{\alpha}},\ddot{\boldsymbol{\alpha}}).

Note that \dot{\boldsymbol{\alpha}}=\mathbf{t} and \ddot{\boldsymbol{\alpha}}=\dot{t} are orthogonal,

\| \boldsymbol{\alpha}\|=1 and \| \ddot{\boldsymbol{\alpha}}\|=\| \dot{\mathbf{t}}\|=\| \kappa\|.

We obtain

\kappa (t)=\displaystyle{\frac{\|\boldsymbol{\alpha}' \wedge\boldsymbol{\alpha}''\|}{\|\boldsymbol{\alpha}'\|^3}}.

iii.

First,

\dot{\boldsymbol{\alpha}}=\displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}s}}=\displaystyle{\frac{\mathrm{d}\boldsymbol{\alpha}}{\mathrm{d}t}}\displaystyle{\frac{\mathrm{d}t}{\mathrm{d}s}}=\boldsymbol{\alpha}'\dot{t};

secondly,

\ddot{\boldsymbol{\alpha}}=\displaystyle{\frac{\mathrm{d}}{\mathrm{d}s}}(\boldsymbol{\alpha}'\dot{t})=\boldsymbol{\alpha}'\ddot{t}+\bigg( \displaystyle{\frac{\mathrm{d}}{\mathrm{d}s}}\boldsymbol{\alpha}' \bigg)\dot{t}=\boldsymbol{\alpha}'\ddot{t}+\boldsymbol{\alpha}''\dot{t}^2;

thirdly,

\dddot{\boldsymbol{\alpha}}=\displaystyle{\frac{\mathrm{d}}{\mathrm{d}s}}\big( \boldsymbol{\alpha}'\ddot{t}+\boldsymbol{\alpha}''\dot{t}^2 \big) =\boldsymbol{\alpha}'\dddot{t}+\boldsymbol{\alpha}''\dot{t}\ddot{t}+\boldsymbol{\alpha}''2\dot{t}\ddot{t}+\boldsymbol{\alpha}'''\dot{t}^3=\boldsymbol{\alpha}'\dddot{t}+2\boldsymbol{\alpha}''\dot{t}\ddot{t}+\boldsymbol{\alpha}'''\dot{t}^3.

Compute \langle \dot{\boldsymbol{\alpha}}\wedge \ddot{\boldsymbol{\alpha}}, \dddot{\boldsymbol{\alpha}} \rangle as follows:

\begin{aligned} \langle \dot{\boldsymbol{\alpha}}\wedge \ddot{\boldsymbol{\alpha}}, \dddot{\boldsymbol{\alpha}} \rangle & = \big( \boldsymbol{\alpha}'\dot{t} \big) \wedge \big( \boldsymbol{\alpha}'\ddot{t}+\boldsymbol{\alpha}''\dot{t}^2 \big) \cdot \big( \boldsymbol{\alpha}'\dddot{t}+2\boldsymbol{\alpha}''\dot{t}\ddot{t}+\boldsymbol{\alpha}'''\dot{t}^3 \big) \\ & =\big( \boldsymbol{\alpha}'\dot{t}\wedge \boldsymbol{\alpha}''\dot{t}^2\big) \cdot \big( \boldsymbol{\alpha}'\dddot{t}+2\boldsymbol{\alpha}''\dot{t}\ddot{t}+\boldsymbol{\alpha}'''\dot{t}^3 \big)\\ \end{aligned}

Now that the cross product \boldsymbol{\alpha}'\wedge \boldsymbol{\alpha}'' is orthogonal to both \boldsymbol{\alpha}' and \boldsymbol{\alpha}'', we can ignore the dot product among them and what remains is

(\boldsymbol{\alpha}'\dot{t}\wedge \boldsymbol{\alpha}''\dot{t}^2)\cdot \boldsymbol{\alpha}'''\dot{t}^3, or,

\dot{t}^6\langle \boldsymbol{\alpha}'\wedge \boldsymbol{\alpha}'', \boldsymbol{\alpha}'''\rangle.

Using the result of part i., substitute 1/\| \boldsymbol{\alpha}'\| for \dot{t}, we obtain

\langle \dot{\boldsymbol{\alpha}}\wedge \ddot{\boldsymbol{\alpha}}, \dddot{\boldsymbol{\alpha}} \rangle=\displaystyle{\frac{\langle \boldsymbol{\alpha}'\wedge \boldsymbol{\alpha}'', \boldsymbol{\alpha}'''\rangle}{\|\boldsymbol{\alpha}'\|^6}}.

Using the formula for curvature in part ii. and put it into

\tau =\displaystyle{\frac{\langle \dot{\boldsymbol{\alpha}}\wedge \ddot{\boldsymbol{\alpha}},\dddot{\boldsymbol{\alpha}} \rangle}{\kappa^2}}=\displaystyle{\frac{\langle \boldsymbol{\alpha}'\wedge \boldsymbol{\alpha}'', \boldsymbol{\alpha}'''\rangle}{\kappa^2 \|\boldsymbol{\alpha}'\|^6}},

i.e.,

\tau (t)=\displaystyle{-\frac{\langle \boldsymbol{\alpha}' \wedge \boldsymbol{\alpha}'',\boldsymbol{\alpha}'''\rangle }{\| \boldsymbol{\alpha}' \wedge\boldsymbol{\alpha}''\|^2}}.