- Calculus, Volume 2: Multi-Variable Calculus and Linear Algebra with Applications to Differential Equations and Probability
- Tom M. Apostol
- Second Edition
- 1991
- 978-1-119-49676-2
2.8 Exercises
31. $\quad$ Let $V$ be the linear space of all real polynomials $p(x).$ Let $R,$ $S,$ $T$ be the functions which map an arbitrary polynomial $p(x) = c_0 + c_1 x + \cdots + c_n x^n$ in $V$ onto the polynomials $r(x),$ $s(x),$ and $t(x),$ respectively, where
\begin{align*}
r(x) = p(0), \quad s(x) = \sum_{k=1}^n c_k x^{k-1}, \quad t(x) = \sum_{k=0}^n c_k x^{k+1}.
\end{align*}
(a) $\quad$ Let $p(x) = 2 + 3x - x^2 + x^3$ and determine the image of $p$ under each of the following transformations: $R,$ $S,$ $T,$ $ST,$ $TS,$ $(TS)^2,$ $T^2 S^2,$ $S^2 T^2,$ $TRS,$ $RST.$
(b) $\quad$ Prove that $R,$ $S,$ and $T$ are linear and determine the null space and range of each.
(c) $\quad$ Prove that $T$ is one-to-one and determine its inverse.
(d) $\quad$ If $n \geq 1,$ express $(TS)^n$ and $S^n T^n$ in terms of $I$ and $R.$
Solution.
(a) $\quad$ Let $p(x) = 2 + 3x - x^2 + x^3$ and determine the image of $p$ under each of the following transformations: $R,$ $S,$ $T,$ $ST,$ $TS,$ $(TS)^2,$ $T^2S^2,$ $S^2 T^2,$ $TRS,$ $RST.$
\begin{align*}
R[p(x)] &= r(x)
\\
&= p(0)
\\
&= 2
\\
S[p(x)] &= s(x)
\\
&= \sum_{k=1}^n c_k x^{k-1}
\\
&= 3 - x + x^2
\\
\\
T[p(x)] &= t(x)
\\
&= \sum_{k=0}^n c_k x^{k+1}
\\
&= 2x + 3x^2 - x^3 + x^4
\\
\\
ST[p(x)] &= S[T[p(x)]]
\\
&= S[2x + 3x^2 - x^3 + x^4]
\\
&= 2 + 3x - x^2 + x^3
\\
\\
TS[p(x)] &= T[S[p(x)]]
\\
&= T[3 - x + x^2]
\\
&= 3x - x^2 + x^3
\\
\\
(TS)^2[p(x)] &= TS[TS[p(x)]]
\\
&= T[S(3x - x^2 + x^3)]
\\
&= T(3 - x + x^2)
\\
&= 3x - x^2 + x^3
\\
\\
T^2S^2[p(x)] &= T^2S[3 - x + x^2]
\\
&= T[T(-1 + x)]
\\
&= T(-x + x^2)
\\
&= -x^2 + x^3
\\
\\
S^2T^2[p(x)] &= S^2T(2x + 3x^2 - x^3 + x^4)
\\
&= S[S(2x^2 + 3x^3 - x^4 + x^5)]
\\
&= S(2x + 3x^2 - x^3 + x^4)
\\
&= 2 + 3x - x^2 + x^3
\\
\\
TRS[p(x)] &= T[R(3 - x + x^2)]
\\
&= T(3)
\\
&= 3x
\\
\\
RST[p(x)] &= R[ST[p(x)]]
\\
&= R(2 + 3x - x^2 + x^3)
\\
&= 2 \quad \blacksquare
\end{align*}
(b) $\quad$ Prove that $R,$ $S,$ and $T$ are linear and determine the null space and range of each.
Proof. $\quad$ Recall the definition of a linear transformation from Section 1:
$\quad$ Definition. $\quad$ If $V$ and $W$ are linear spaces , a function $T: V \to W$ is called a linear transformation of $V$ into $W$ if it has the following two properties:
$\quad$ (a) $\quad T(x + y) = T(x) + T(y) \quad$ for all $x$ and $y$ in $V,$
$\quad$ (b) $\quad T(cx) = cT(x) \quad$ for all $x$ in $V$ and all scalars $c.$
Thus, we wish to show that $R,$ $S,$ and $T$ satisfy these two properties. Let $p$ and $q$ be two polynomials in $V$ given by:
\begin{align*}
p(x) &= \sum_{k=0}^n c_kx^k, \quad q(x) = \sum_{k=0}^n d_kx^k
\end{align*}
and let $a$ and $b$ be scalar constants. We then have:
\begin{align*}
R[ap(x) + bq(x)] &= R\left[\sum_{k=0}^n (ac_k + bd_k)x^k\right]
\\
&= ac_0 + bd_0
\\
&= ap(0) + bq(0)
\\
&= aR[p(x)] + bR[q(x)]
\\
\\
S[ap(x) + bq(x)] &= S\left[\sum_{k=0}^n (ac_k + bd_k)x^k\right]
\\
&= \sum_{k=1}^n (ac_k + bd_k)x^{k-1}
\\
&= a\sum_{k=1}^n c_kx^{k-1} + b\sum_{k=1}^n d_kx^{k-1}
\\
&= aS[p(x)] + bS[q(x)]
\\
\\
T[ap(x) + bq(x)] &= T\left[\sum_{k=0}^n (ac_k + bd_k)x^k\right]
\\
&= \sum_{k=0}^n (ac_k + bd_k)x^{k+1}
\\
&= a\sum_{k=0}^n c_kx^{k+1} + b\sum_{k=0}^n d_kx^{k+1}
\\
&= aT[p(x)] + bT[q(x)]
\end{align*}
As we can see, $R,$ $S,$ and $T$ are linear transformations. Their null spaces and ranges are given as follows:
$\quad$ $N(R)$ is the set of all polynomials in $V$ with $c_0 = 0.$ $R(V)$ is the set of all constant polynomials.
$\quad$ $N(S)$ is the set of all polynomials in $V$ with coefficients $c_1, \dots, c_n$ equal to zero, and arbitrary $c_0.$ In other words, $N(S)$ is the set of all constant polynomials. $S(V) = V.$
$\quad$ $N(T)$ contains only the zero polynomial. $T(V)$ is the set of all polynomials $p$ such that $p(0) = 0. \quad \blacksquare$
(c) $\quad$ Prove that $T$ is one-to-one and determine its inverse.
$\quad$ Proof. $\quad$ By definition, if $T$ is one-to-one on $V,$ then for any two polynomials $p$ and $q$ in $V,$ $p \neq q$ implies $T(p) \neq T(q).$ Now, let $p$ and $q$ be elements of $V$ given by
\begin{align*}
p(x) &= \sum_{k=0}^n c_kx^k, \quad q(x) = \sum_{k=0}^n d_kx^k
\end{align*}
where for some $i,$ in $1, \dots, n,$ $c_i \neq d_i.$ In other words, $p \neq q.$ Taking their respective transformations, we get
\begin{align*}
Tp(x) &= \sum_{k=0}^n c_kx^{k+1}, \quad Tq(x) = \sum_{k=0}^n d_kx^{k+1}
\end{align*}
But since $c_i \neq d_i,$ $Tp(x) \neq Tq(x).$ Hence, $p \neq q$ implies $Tp \neq Tq,$ making $T$ one-to-one on $V.$
$\quad$ The inverse of $T$ is the transformation $T^{-1}p$ that decreases the power of $x$ by $1$ for all $x^k$ where $k \geq 1.$ Namely, the inverse of $T$ is $S. \quad \blacksquare$
(d) $\quad$ If $n \geq 1,$ express $(TS)^n$ and $S^n T^n$ in terms of $I$ and $R.$
$\quad$ Let $p(x) = \sum_{k=0}^n c_kx^k$ be an arbitrary polynomial in $V.$ From part (c), we know that $S = T^{-1},$ giving us $ST = I_V.$ Applying $S$ as a left-inverse to $T,$ we get $ST = I_V,$ and $(TS)^n$ becomes
\begin{align*}
(TS)^n[p(x)] &= T(ST)^{n-1}S[p(x)]
\\
&= T(I^{n-1})S[p(x)]
\\
&= TS\left[\sum_{k=0}^n c_kx^k\right]
\\
&= T\left[\sum_{k=1}^n c_kx^{k-1}\right]
\\
&= \sum_{k=1}^n c_kx^k
\\
&= p(x) - p(0)
\end{align*}
But by definition, we have $p(x) - p(0) = I[p(x)] - R[p(x)].$ Then, by linearity, we can rewrite $I[p(x)] - R[p(x)] = (I - R)[p(x)]$ to give us:
\begin{align*}
(TS)^n &= I - R
\end{align*}
$\quad$ To express $S^nT^n$ in terms of $I$ and $R,$ we recall that $ST = I_V.$ Hence, we can use the associative property of composition and the result of Exercise 21 to give us
\begin{align*}
S^nT^n &= S^{n-1}(ST)T^{n-1}
\\
&= S^{n-1}(I)T^{n-1}
\\
&= S^{n-2}(ST)T^{n-2}
\\
&\dots
\\
&= ST
\\
&= I \quad \blacksquare
\end{align*}