3 4

1
ANSWER:- Now, {𝑥n𝑘 } is a subsequence of {xn}. Set y0 = 𝑥n1 and yk = 𝑥n𝑘+1 − 𝑥n𝑘 ∀, k ≥ 1. Note that kykk ‖𝑦𝑘 ‖= ‖𝑥n𝑘+1 − 𝑥n𝑘 ‖< and
2𝑘
(e) If x and y are two vectors in a Hilbert space then (𝑥, 𝑦) = ‖𝑥‖‖𝑦‖ 1
ANSWER:- ∑𝑘𝑖=1 𝑦𝑖 = 𝑥n𝑘+1 . Now ∑∞ ∞
𝑘=0‖𝑦𝑘 ‖ = ‖𝑦0 ‖ + ∑𝑘=1‖𝑦𝑘 ‖ < ‖𝑦0 ‖ + 2𝑘 = ‖𝑦0 ‖ + 1 < ∞ ,
(f) if (𝒆𝟏 , 𝒆𝟐 , ⋯ , 𝒆𝒏 ) be a finite orthonormal set in a Hilbert spaces H and x is any vector in H, then which one of the following is correct: Thus {𝑦𝑘 } is absolutely summable and hence it summable by assumption. Hence ∑∞𝑘=0‖𝑦𝑘 ‖= x ∈ X.
(i) 𝐱 − ∑𝒏𝒊=𝟏(𝒙, 𝒆𝒊 )𝒆𝒊 is perpendicular to each 𝒆𝒊 (ii) TT ′ = T ′ T is perpendicular to each 𝑒𝑗 , (iii) x − ∑𝑛𝑖=1(𝑥, 𝑒𝑖 )𝑒𝑖 is equal to 𝑒𝑗 for each 𝑖 Now lim 𝑥n𝑘+1 = lim ∑𝑘𝑖=1 𝑦𝑖 = ∑∞ 𝑖=0 𝑦𝑖 = 𝑥 Thus lim 𝑥k = x ∈ X. Thus, the Cauchy sequence {xn} has a convergent
ANSWER:- Since if {e1 , e2 , … , en } be a finite orthonormal in a Hilbert space H and x be any vector in H then x − ∑ni=1(x, ei ) ei ⊥ ej for ech j 𝑘→∞ 𝑘→∞ 𝑘→∞
subsequence {𝑥n𝑘 } convergent to x. Therefore lim 𝑥𝑛 = x ∈ X. Hence X is a Banach space.
𝑘→∞
(g) An operator T on a Hilbert spaces H is said to be an unitary operator iff TT ∗ = T ∗ T = I, where T ∗ and I have their usual meanings 2018-3. Let N be a non-zero normed linear subspace and let S = {𝐱 ∈ 𝐍: ‖𝐱‖ ≤ 𝟏} be a linear space of N . Prove that N is a Banach space 
ANSWER:- TRUE S is complete.
----------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------- PROOF:- 1ST PART:- We assume that N is a Banach space and 𝐒 = {𝐱 ∈ 𝐍 ∶ ‖𝐱‖ ≤ 𝟏}be a linear subspace of N, Then we shall show that S is
2016-1. Each question has multiple answers in which one is correct select the correct answer:- complete:- Let 〈xn 〉 be the any Cauchy Sequence in S,  ‖xn ‖ ≤ 1 ,  n  N, {by definition of S}
2016-1(a) In a Normed linear spaces E which one is correct:- (i) ‖x‖ − ‖y‖ ≠ ‖x − y‖,x, yE, (ii) ‖x‖ − ‖y‖ ≥ ‖x − y‖,x, yE, Since S  N,  〈xn 〉 ∈ N, i.e. 〈xn 〉 is also a Cauchy Sequence in N.
(iii) ‖𝐱‖ − ‖𝐲‖ ≤ ‖𝐱 − 𝐲‖,x, yE. Again Since N is complete being a Banach Space   a point x  N, s. t. xn → x
ANSWER:- (iii) ‖x‖ − ‖y‖ ≤ ‖x − y‖,x, yE Since norm is a continuous function xn → x  ‖xn ‖ → ‖x‖, i. e. ‖x‖ = lim {‖xn ‖} ≤ 14, n, {by definition of S}
n→∞
 x  S, In this way we see that every Cauchy sequence in S converges in S. This proves 1st part of the equation.
2016-1(b) If B and B be Banach spaces and T a linear transformation from B into B then T is continuous iff graph is:- (i) semi-open, (ii) open,
CONVERSE PART:- We assume that the linear subspace S is complete and then we shall show that N is a Banach space:- Since N is a Banach space
(iii) semi-closed, (iv) closed
so we need only to prove that N is complete for this we shall show that every Cauchy sequence in N converges in N :- Let 〈yn 〉 be the any Cauchy
ANSWER:- (iv) closed
Sequence in N,  ‖ym − yn ‖ → 0 ,  m, n   , . . . (I)
y 1
We define n , xn = ‖yn ‖ ,  ‖xn ‖ = ‖yn ‖ = 1,  xn ∈ S , {by definition of S}
2016-1(c) In an inner product spaces E if (x|y) denotes the inner product of x with y for any pair (x, y)E x E then (x|y + z) his equal to:- (i) n ‖yn ‖

(x|y) + (x|z), (ii) (y|x) + (z|x), (iii) ̅ (𝐱|𝐲) + ̅(𝐱|𝐳) Now we shall show that 〈xn 〉 is a Cauchy sequence in S:- we have ‖xm − xn ‖ = ‖‖ym ‖ − ‖yn ‖‖
y y
m n
ANSWER:- (iii) ̅ (𝐱|𝐲) + ̅(𝐱|𝐳) y y y y y y y y
= ‖(‖ym‖ − ‖y n ‖) + (‖y n ‖ − ‖yn ‖)‖ ≤ ‖(‖ym‖ − ‖y n ‖)‖ + ‖‖y n ‖ − ‖yn ‖‖ , {by prop. Of norm}
m m m n m m m n
ym −yn 1 1 1 1 1
2016-1(d) In an inner product spaces E if ‖x‖ = +√x|x then which one of the following is correct:- (i) ‖x‖ − |x|y| ≠ ‖x‖‖y‖,x, yE, =‖ ‖ym ‖
‖ + ‖yn (‖y − )‖ = ‖‖y ‖ ‖ym − yn ‖ + ‖(‖y − ‖y ‖)‖ ‖yn ‖
m‖ ‖yn ‖ m‖ m‖ n
(ii) |𝐱|𝐲| ≤ ‖𝐱‖‖𝐲‖,x, yE, (iii) |x|x| ≥ ‖x‖‖y‖,x, yE. 1 | ‖yn ‖−‖ym ‖ | 1 1
= ‖y ‖ym − yn ‖ + ≤ ‖y ‖ym − yn ‖ + ‖ym − yn ‖
ANSWER:- (ii) |x|y| ≤ ‖x‖‖y‖,x, yE m‖ ‖ym ‖ m‖ ‖ym ‖
1
Thus we get ‖xm − xn ‖ ≤ 2 ‖y ‖ym − yn ‖ ,  0,  m , n   , { by (I) }
m‖
2016-1(e) For any two elements x and y in a Hilbert spaces H which one of the following is correct? ‖x + y‖2 + ‖x − y‖2 is equal to :- Hence sequence 〈xn 〉 is a Cauchy sequence in S, Since S is complete   xS, s. t. xn  x , 
yn
→ x , ....(II)
‖yn ‖
(i) ‖x‖2 + ‖y‖2 , (ii) ‖x‖2 + 2‖y‖2 , (iii) 𝟐‖𝐱‖𝟐 + 𝟐‖𝐲‖𝟐 ,
ANSWER:- (iii) 2‖x‖2 + 2‖y‖2 Also we have | ‖yn ‖ − ‖ym ‖ | ≤ ‖yn − ym ‖ ,  0,  m , n   , {by (I)}
 〈‖yn ‖〉 is a Cauchy sequence of real numbers in R
2016-1(f) if (e1 , e2 , ⋯ , en ) be a finite orthonormal set in a Hilbert spaces H and x is any vector in H, then which one of the following is correct: Since R is complete  〈‖yn ‖〉 → α ∈ R , (say),  by (II) yn → αx ∈ N . Hence N is complete. This completes the proof of the theorem.
(i) 𝐱 − ∑𝐧𝐢=𝟏(𝐱, 𝐞𝐢 )𝐞𝐢 is perpendicular to each 𝐞𝐢 (ii) TT ′ = T′T is perpendicular to each ej , (iii) x − ∑ni=1(x, ei )ei is equal to ej for each i
ANSWER:- (i) 𝐱 − ∑𝐧𝐢=𝟏(𝐱, 𝐞𝐢 )𝐞𝐢 is perpendicular to each 𝐞𝐢 2018-4. Let N and N be Normed linear spaces and B(N, N) denotes the set of all bounded (or continuous) linear Transformations from N into
N’ then B(N, N) is itself a normed linear space with respect to point wise linear operations: (𝐓 + 𝐔)𝐱 = 𝐓(𝐱) + 𝐔(𝐱),
2016-1(g) An operator T on a Hilbert spaces H is said to be an unitary operator which one is correct?:(i) T T  = T ′ T, (ii) 𝐓𝐓 ′ = 𝐓 ′ 𝐓 = 𝐈, (𝐓)(𝐱) = 𝐓(𝐱) and the norm defined by ‖𝐓‖ = 𝐬𝐮𝐩. {‖𝐓(𝐱)‖: 𝐱 ∈ 𝐍: ‖𝐱‖ ≤ 𝟏}, further if N is a Banach Space then so is
(iii) TT ′ = T ′ T = T B(N,N) P – 55
ANSWER:- (ii) (ii) 𝐓𝐓 ′ = 𝐓 ′ 𝐓 = 𝐈 ANSWER:- STEP – I: 1st We shall show that B(N, N’) is a Linear Space:- We know that the det of linear transformations from one linear space to
______________________________________________________________________________________________________________________ another linear space is a linear space with respect to given definition of addition and scalar multiplication of linear transformations:- i.e. xN,
F, ∀ T, U ∈ B(N, N ′ ), (T + U)x = T(x) + U(x), (T)(x) = T(x), It will be therefore sufficient if we prove that the set of all
MSC – 10 – LONG ANSWER TYPE QUESTION PAPERS bounded (or continuous) linear Transformations from N into N’ is a subspace of the space of Linear transformations:- By criteria of Subspace
2018-2. A normed linear space N is a Banach space if every absolutely summable series in N is summable. P – 213 we know that The non-empty subset W of Linear Space V(F) is a subspace is that ∀ α, β ∈ W, ∀a, b ∈ F  aα + bβ ∈ W,
OR A normed space X is a Banach space iff every absolutely summable sequence in X is summable in X. Let T, U ∈ B(N, N ′ ), then by boundedness of L.T. we have ‖T(x)‖ ≤ k1 ‖x‖, ‖U(x)‖ ≤ k 2, ‖x‖, xN and for some +ve real k1, k 2, }, (i)
Definition: Let N be a normed space over the field F. A series ∑ fn ⊂ N is said to be summable to a sum s isf s is in the space Now we have ‖(aT + bU)x‖ = ‖aT(x) + bU(x)‖, {by definition of addition in B(N, N ′ ) }
and the sequence of partial sums of the series converges to s i.e., let ‖s − ∑nk=1 fk ‖ as k  ∞. Then we write s = ∑∞  ‖(aT + bU)x‖ ≤ ‖aT(x)‖ + ‖bU(x)‖ , {∵ N ′ is a normed linear space}
k=1 fk
 ‖(aT + bU)x‖ ≤ |a|‖T(x)‖ + |b|‖U(x)‖ , {∵ N ′ is a normed linear space}
The series ∑∞ ∞
k=1 fk is said to be absolutely convergent if ∑k=1‖fk ‖ < ∞  ‖(aT + bU)x‖ ≤ |a|k1 ‖x‖ + |b|k 2 ‖x‖ , { Using (i)}
 ‖(aT + bU)x‖ ≤ (|a|k1 + |b|k 2 )‖x‖ , { ∵ N is a normed linear space}
PROOF: If Part (⇒) Suppose that X is a Banach space. Let {xn} be an absolutely summable sequence in X. ∑∞ n=1‖xn ‖ = M < ∞
 ‖(aT + bU)x‖ ≤ K‖x‖ , { where K = |a|k1 + |b|k 2 ∈ F}
Hence for each ε > 0, ∃ N ∈ N such that ∑∞n=N‖xn ‖ < ε. Now, if n ≥ m > N  aT + bU is bounded. i.e. continuous and hence belongs to B(N, N ′ ), Consequently B(N, N ′ ) is a subspace
⇒ ‖sn − 𝑠𝑚 ‖ = ‖∑nk=m+1 xk ‖ ≤ ∑nk=m+1‖xk ‖ ≤ ∑∞ k=N k‖x ‖ < ε STEP – II: Now We shall show that B(N, N’) is a normed Linear Space:- Let T ∈ B(N, N ′ ), then by given definition of norm we have
Thus {sn} is a Cauchy sequence in X, hence {sn} is convergent since X is Banach space. Therefore {xn} is summable. ‖T‖ = sup. {‖T(x)‖: x ∈ N: ‖x‖ ≤ 1}, Now we have to verify all the four axioms of norm in B(N, N ′ ):-
Convese Part (⇐) Suppose each absolutely summable sequence in X is summable in X. Let {xn} ⊂ X be a Cauchy sequence in X. Now, [N1] :- Since ‖T(x)‖ ≥ 0,xN , {∵ N ′ is a normed linear space}
1
since {xn} is Cauchy, ∃ n1 ∈ N such that if n, m ≥ n1 ⇒‖𝑥n − 𝑥m ‖ < ½ . Also, ∃ n2’ ∈ N such that if n, m ≥ n2’ ⇒‖𝑥n − 𝑥m ‖ < 22 and  ‖T‖ ≥ 0, Thus [N1} holds
1 1 [N2] :- Since ‖T‖ = sup. {‖T(x)‖: x ∈ N: ‖x‖ ≤ 1} , {by given definition of norm }
let n2 > max{n1,n2’ }. Now, n2 > n1 and if n, m ≥ n2 ⇒ ‖𝑥n − 𝑥m ‖< . Hence n1, n2 ≥ n2 ⇒ ‖𝑥n2 − 𝑥n1 ‖ < . Continuing this way, we ‖T(x)‖
22 22 = sup. { : x ∈ N, 0 ≠ x}
1 ‖x‖
have for each k ≥ 2 ∃n𝑘+1 > n𝑘 such that ‖𝑥n𝑘+1 − 𝑥n𝑘 ‖<2𝑘 ‖T(x)‖
 ‖T‖ = 0  ‖x‖
= 0, xN and x≠ 0

7 8

‖I(x)‖ = ‖x‖ ≤ ‖x‖ + ‖T(x)‖,  ‖I(x)‖ ≤ ‖x‖1 , { as ‖x‖1 = ‖x‖ + ‖T(x)‖} f(x) f(x)
Now we can write f(x) = )
f(y0 ) = βf(y0 ) , where )
= β, Thus we have f(x) = βf(y0 ), f(x) − βf(y0 ) = 0,
f(y0 f(y0
This shows that I is bounded and hence continuous.
 f(x − βy0 ) = 0, {as f is linear}
To show that I is a homeomorphism it is enough to show that B1 is complete:- Let 〈xn 〉 be the any Cauchy sequence in B1 So that
 x − βy0 ∈ 𝑀, which is the null space of f
‖xn − xm ‖ → 0 ,  m, n → ∞,  ‖xn − xm ‖ + ‖T(xn − xm )‖ → 0 ,  m, n → ∞,  ‖xn − xm ‖ + ‖T(xn ) − T(xm )‖ → 0 ,  m, n → ∞
 x − βy0 = 𝑚 for some m  M
 ‖xn − xm ‖ → 0 and ‖T(xn ) − T(xm )‖ → 0 ,  m, n → ∞
Thus x  H  x = m + y0 , Now f(x) = f(m + y0 ) = f(m) + f(y0 ) , {as f is linear}
This shows that if 〈xn 〉 is a Cauchy sequence in B1 then 〈T(xn )〉 is a Cauchy sequence in B’
 f(x) = (m , y) + (y0 , 𝑦) , {by assumption f(x) = (x , y) }
Again since B and B’ are complete  we have xn → x ∈ B , T(xn ) → y ∈ B′ , . . . (i)
 f(x) = (m , y) + (y0 , 𝑦) , {by property of inner product }
Since graph of T0 is given to be closed  (i) shows that (x, y) ∈ T0  y = T(x)
Now we have ‖xn − x‖1 = ‖xn − x‖ + ‖T(xn − x)‖ {by defined definition of ‖ ‖1 }  f(x) = (𝑚 + y0 , 𝑦) , {by property of inner product }
 ‖xn − x‖1 = ‖xn − x‖ + ‖T(xn ) − T(x)‖  ‖xn − x‖1 = ‖xn − x‖ + ‖T(xn ) − y‖  ‖xn − x‖1 → 0, { as xn → x , T(xn ) → y }  f(x) = (𝑥, 𝑦) , {as 𝑥 = 𝑚 + y0 }
 Sequence 〈xn 〉 in B1 converges to xB1 and hence B1 is complete. This completes the proof of the theorem. Thus we conclude that if a vector if y satisfies (i)  x  M and for x = y0 then it satisfies (i) xH. Hence 𝑦 = αy0 is the required vector,
̅̅̅̅̅̅̅
f(y 0)
where  = ‖y 2 This completes the proof of the theorem.
0‖
2018-7. Prove that a closed convex subset C of a Hilbert space H contains a unique vector of smallest norm. P – 130
PROOF:- 1st we shall define a convex set in a Hilbert space:- A non-empty subset S of a Hilbert space H is said to be a Convex set if x, y  S 2018-9 The adjoint operator T T* on B(H) then prove that: (a) (𝐓𝟏 + 𝐓𝟐 )∗ = 𝐓𝟏 ∗ + 𝐓𝟐 ∗ , (b) (𝐓𝟏 𝐓𝟐 )∗ = 𝐓𝟐 ∗ 𝐓𝟏 ∗ , (c) ‖𝐓 ∗ ‖ = ‖𝐓‖,
x+y
 x + (1 - )  S, where 0    1, for convenience we take  = ½ . Thus S is a convex set if x, y  S  ∈ S, (d) ‖𝐓 ∗ 𝐓‖ = ‖𝐓‖𝟐 , (e) 𝐓 ∗∗ = 𝐓
2
Proof of the theorem:- Let C be a closed convex subset of a Hilbert space H. Let d = inf. {‖x‖: xC}, then there must exist a sequence 〈xn 〉 of ANSWER:- SEE Q. No. 8 of MSC 10 -2016
x𝑚+xn
vectors in C, such that ‖xn ‖ → d. Let any two vectors x𝑚 , xn belongs to 〈xn 〉. Since C ius a convex subset of H  x𝑚 , xn ∈ C,  ∈C ________________________________________________________________________________________________________________________
2
x𝑚 +xn 2017-2. Define Normed linear space Prove that in a Normed linear space E (i) ‖𝒙‖ − ‖𝒚‖ ≤ ‖𝒙 − 𝒚‖,x, yE, (ii) for every sequence {𝐱 𝐧 } in
Hence by definition of d we have ‖ ‖ ≥ d  ‖x𝑚 + xn ‖ ≥ 2d , ... (i)
2 E, xE 𝐱 𝐧 → 𝐱  ‖𝐱 𝐧 ‖ → ‖𝐱‖
In H by Parallelogram law we have ‖x𝑚 + xn ‖2 + ‖x𝑚 − xn ‖2 = 2‖x𝑚 ‖2 + 2‖xn ‖2 ,  ‖x𝑚 − xn ‖2 = 2‖x𝑚 ‖2 + 2‖xn ‖2 − ‖x𝑚 + xn ‖2 ANSWER:- Definition of a Normed linear space :- Let N be a complex or Real linear space Let ‖ ‖ be a function from N into R, s.t.  x , y  N and
 ‖x𝑚 − xn ‖2 ≤ 2‖x𝑚 ‖2 + 2‖xn ‖2 − 4d2 , ... (ii), {by (i)}    C or R , we have:- [N1]: ‖x‖ ≥ 0, [N2]: ‖x‖ = 0 ⇔ x = 0, [N3]: ‖x + y‖ ≤ ‖x‖ + ‖y‖, [N4]: ‖αx‖ ≤ |α| ‖x‖,
Now since ‖xm ‖ → d , ‖xn ‖ → d,  By (ii) we have 2‖x𝑚 ‖2 + 2‖xn ‖2 − 4d2 → 2d2 + 2d2 − 4d2 → 0 Then the function ‖ ‖, is called a norm on N and the pair (N, ‖ ‖) is called a Normed linear space.
Hence as m , n   we have ‖x𝑚 − xn ‖2 → 0, i.e. ‖x𝑚 − xn ‖ → 0, consequently 〈xn 〉. is a Cauchy sequence in C.  〈xn 〉 is a Convergent (i) Let E be a normed linear space N and let x, y  E then | ‖x‖ − ‖y‖ | ≤ ‖x − y‖
sequence and converges in C. Let x  C, s. t. xn → x , i.e. x = lim. xn  ‖x‖ = lim. ‖xn ‖ = d. Hence x is a vector in C with smallest norm. PROOF:- Let E be a normed linear space also let x, y  E, Now we have
Uniqueness of x:- Let x and x are two vectors in C, such that ‖x‖ = d , ‖x′‖ = d then we have to show that x = x:- ‖x‖ = ‖(x − y) + y‖ ≤ ‖(x − y)‖ + ‖y‖, {by [N3] in normed linear space E}
x+x′ x+x′
Since C is a convex set x, x′ ∈ C 
2
∈ C. Hence by the definition of d we have
2
≥ d  x + x′ ≥ 2d  ‖x‖ − ‖y‖ ≤ ‖(x − y)‖, … (i) This Proves (i)
Now by Parallelogram law we have ‖𝑥 − x′‖2 = 2‖x‖2 + 2‖x′‖2 − ‖𝑥 + x′‖2 , Again we have ‖y‖ = ‖(y − x) + x‖
≤ 2‖x‖2 + 2‖x′‖2 − 4d2 , {as ‖𝑥 + x′‖2 ≥ 4d2 } ≤ ‖−(x − y)‖ + ‖x‖, {by [N3] in normed linear space E}
→ 2d2 + 2d2 − 4d2 , {as ‖𝑥‖ → d , ‖𝑥′‖ → d,}  ‖y‖ − ‖x‖ ≤ |−1|‖(x − y)‖{by [N4] in normed linear space E}
→0,  ‖y‖ − ‖x‖ ≤ ‖(x − y)‖,
 ‖𝑥 − x′‖2 → 0 ,  ‖𝑥 − x′‖ → 0,  𝑥 = x′, This completes the proof of the theorem.  −(‖x‖ − ‖y‖) ≤ ‖(x − y)‖, … (ii)
𝑥, 𝑥 ≥ 0
Combining (i) and (ii) and using definition of mod function i.e. |𝑥| = ⟦ , we get | ‖x‖ − ‖y‖ | ≤ ‖x − y‖ PROVED.
2018-8. Let H be a Hilbert space and f be an arbitrary function in H* then there exist a unique vector y in H such that f(x) = (x, y), xH, −𝑥, 𝑥 < 0
Prove it. OR State and prove Reisz’s representation theorem OR Reisz representation theorem for continuous linear functionals on a (ii) for every sequence {𝐱 𝐧 } in E, xE 𝐱 𝐧 → 𝐱  ‖𝐱 𝐧 ‖ → ‖𝐱‖
Hilbert space P-356 PROOF:- Let 𝑥 ∈ 𝑁 and let 〈xn 〉 be a sequence in N such that xn → 𝑥 as 𝑛 → ∞, ... (1)
Now we have |f(xn ) − f(x)| = |‖xn ‖ − ‖x‖| ≤ ‖xn − x‖ , {by (i)}
PROOF:- 1st we shall show that if  a vector y in H such that f(x) = f(x, y) xH, then y is necessarily unique :- If possible let y is not unique. Let y1 and
→ 0 as n → ∞, ..., {by (1)}
y2 be the two vectors in H which satisfy the property (i), then we have xH, f(x) = (x, y1 ) also f(x) = (x, y2 ) these results implies that
Hence f(xn ) → f(x) , as 𝑛 → ∞ consequently f is continuous i.e. norm function is continuous. This completes the proof of the theorem.
(x, y1 ) = (x, y2 ), xH,  (x, y1 ) − (x, y2 ) = 0, xH,  (x, y1 − y2 ) = 0, xH, {on using the property of inner product}
 (y1 − y2 , y1 − y2 ) = 0, xH , {on taking 𝑥 = y1 − y2 }
2014-17-3. Define a Banach space and prove that a non-zero normed linear space N is a Banach space iff the set {𝐱 ∈ 𝐍: ‖𝐱‖ = 𝟏} is
 y1 − y2 = 𝟎, xH , {on using the property of inner product }
complete.]
 y1 = y2 , Thus y is unique. PROOF:- 1ST PART:- We assume that N is a Banach space and 𝐒 = {𝐱 ∈ 𝐍 ∶ ‖𝐱‖ ≤ 𝟏}be a linear subspace of N, Then we shall show that S is
Now we shall show that  a vector y in H satisfying (i):- If f is zero functional, then f(x) = 𝟎, xH complete:-
Also if y = 0 then (x, 𝑦) = (x, 0) = 0, xH Let 〈xn 〉 be the any Cauchy Sequence in S,  ‖xn ‖ ≤ 1 ,  n  N, {by definition of S}
Thus If f is zero functional then the vector y = 0 is such that f(x) = (x, y) xH Since S  N,  〈xn 〉 ∈ N, i.e. 〈xn 〉 is also a Cauchy Sequence in N.
Now assuming that f is not a zero functional. i.e. f(x)  0, for some x  H. Let M be the null space of f, i.e. Again Since N is complete being a Banach Space   a point x  N, s. t. xn → x
M = {x ∈ H: f(x) = 0}, then M is a proper subspace of H. Also we know that the null space of any continuous linear transformation is closed. Since f
Since norm is a continuous function xn → x  ‖xn ‖ → ‖x‖, i. e. ‖x‖ = lim {‖xn ‖} ≤ 14, n, {by definition of S}
is continuous  M is a proper closed subspace of H.  by a property of orthogonal set “If M is a proper closed subspace of H, then  a non-zero vector n→∞
 x  S, In this way we see that every Cauchy sequence in S converges in S. This proves 1st part of the equation.
𝑧0 in H s.t. 𝑧0 ⊥ M”. Hence  a non-zero vector y0 ∈ H, s. t. y0 ⊥ M  y0 ∈ M ⊥
CONVERSE PART:- We assume that the linear subspace S is complete and then we shall show that N is a Banach space:- Since N is a Banach space
We shall show that for some suitably chosen scalar  the vector 𝑦 = αy0 will serve our purpose. 1st we observe that whatever may be the value of the
so we need only to prove that N is complete for this we shall show that every Cauchy sequence in N converges in N :- Let 〈yn 〉 be the any Cauchy
scalar  the vector 𝑦 = αy0 satisfy the property (i)  x M which is the null space of f, then f(x) = 0, Also if x M then
Sequence in N,  ‖ym − yn ‖ → 0 ,  m, n   , . . . (I)
(x, y) = (x, y0 ) =  ̅ (x, y0 ) = 0, because x M and y0 ⊥ M, Thus if x M and 𝑦 = αy0 , then we have f(x) = (x, 𝑦) = 𝟎, thus (i) is satisfied. y 1
f(y0 ) We define n , xn = ‖yn ‖ ,  ‖xn ‖ = ‖yn ‖ = 1,  xn ∈ S , {by definition of S}
Now we chose a  s.t. the vector 𝑦 = αy0 satisfies (i) for x = y0 , then f(y0 ) = (y0 , y0 ) =  ̅ ‖y0 ‖2 Thus if we take 
̅ (y0 , y0 ) =  ̅ = ‖y 2
n ‖yn ‖
0‖ y y
̅̅̅̅̅̅̅
f(y0 ) Now we shall show that 〈xn 〉 is a Cauchy sequence in S:- we have ‖xm − xn ‖ = ‖‖ym ‖ − ‖yn ‖‖
or take  = ‖y ‖2
, then the vector 𝑦 = αy0 satisfies (i) for every x in M and for x = y0 m n
0 y y y y y y y y
= ‖(‖ym‖ − ‖y n ‖) + (‖y n ‖ − ‖yn ‖)‖ ≤ ‖(‖ym‖ − ‖y n ‖)‖ + ‖‖y n ‖ − ‖yn ‖‖ , {by prop. Of norm}
Now we shall complete the proof by showing that if a vector y satisfies (i) for every x in M and for x = y0 then it will satisfy (i) for every x in H:- Since m m m n m m m n
ym −yn 1 1 1 1 1
M ∩ M ⊥ = {𝟎} and y0 is a non-zero vector belonging to M ⊥ y0 ∉ M so we have f(y0 ) ≠ 0, =‖ ‖ + ‖yn (‖y − ‖y ‖)‖ = ‖‖y ‖ ‖ym − yn ‖ + ‖(‖y − ‖y ‖)‖ ‖yn ‖
‖ym ‖ m‖ n m‖ m‖ n
 5 6

 ‖T(x)‖ = 0, xN and x≠ 0, {∵ x≠ 0  ‖x‖ ≠ 0} [N1]: By defined definition of norm on R or C we have ‖x‖ = |x|, ∀x ∈ R or C,  ‖x‖ ≥ 0, {by definition of mod function: |x| ≥ 0}, Thus
 T(x) = 0, xN and x≠ 0, {∵ N ′ is a normed linear space} [N1] holds
 T = O, i.e. T is a zero linear transformation. Thus [N2} holds [N2]: By defined definition of norm on R or C we have ‖x‖ = |x|, ∀x ∈ R or C,  ‖x‖ = 0,  |x| = 0  x = 0 , Thus [N2] holds
[N3] :- We have ‖(αT)x‖ = ‖αT(x)‖ , {∵ N ′ is a normed linear space} [N3]: By defined definition of norm on R or C we have ‖x‖ = |x|, ∀x ∈ R or C,  ‖αx‖ = |αx| = |α||x|, , {by propertyof mod function} Thus
= |α|‖T(x)‖ , {∵ N ′ is a normed linear space}  ‖αx‖ = |α|‖x‖ , {by definition of mod function} [N3] holds
Now ‖αT‖ = sup. {‖αT(x)‖: x ∈ N: ‖x‖ ≤ 1}, {by given definition of norm} [N4]: By defined definition of norm on R, we have ‖x + y‖ = |x + y|, ∀x, y ∈ R ‖x + y‖ ≤ |x| + |𝑦|, {by propertyof mod function}
= sup. {|𝛼|‖T(x)‖: x ∈ N: ‖x‖ ≤ 1} , {∵ N ′ is a normed linear space} ‖x + y‖ ≤ ‖x‖ + ‖𝑦‖, , {by definition of mod function}, Thus [N4] holds
= |𝛼|sup. {‖T(x)‖: x ∈ N: ‖x‖ ≤ 1} Again By defined definition of norm on C, we have ‖x + y‖ = |x + y|, ∀x, yC ‖x + y‖2 = |x + y|2 = (x + y)(x̅ + y̅) = xx̅ + yy̅ + xy̅ + x̅y
= |𝛼|‖T‖ , {by given definition of norm}. Thus [N3} holds  ‖x + y‖2 ≤ |x|2 + |y|2 + 2|x̅y| , {note that |xy̅| = |x̅y|} ≤ xy̅ or x̅y}
[N4] :- We have ‖T + U‖ = sup. {‖(T + U)x‖: x ∈ N: ‖x‖ ≤ 1} , {by given definition of norm }  ‖x + y‖2 ≤ |x|2 + |y|2 + 2|x||y| , {note that |xy̅| = |x||𝑦̅| = |x||𝑦|}
= sup. {‖T(x) + U(x)‖: x ∈ N: ‖x‖ ≤ 1} , {∵ N ′ is a normed linear space}  ‖x + y‖2 ≤ (|x| + |y|)2 , {note that |xy̅| = |x||𝑦̅| = |x||𝑦|}
≤ sup. {‖T(x)‖ + ‖U(x)‖: x ∈ N: ‖x‖ ≤ 1} , {∵ N ′ is a normed linear space}  ‖x + y‖ ≤ |x| + |y| ,  ‖x + y‖ ≤ ‖x‖ + ‖y‖, , {by definition of norm function}. Thus [N4] holds.
= sup. {‖T(x)‖: x ∈ N: ‖x‖ ≤ 1} + sup. {‖U(x)‖: x ∈ N: ‖x‖ ≤ 1} In this way all the four axioms of a norm function defined on R or C are satisfied. Hence both R and C are normed linear spaces.
= ‖T‖ + ‖U‖ , {by given definition of norm } Also we know that every normed linear spaces are metric space and each metric spaces are complete spaces Consequently R or C are complete
Thus we get ‖T + U‖ ≤ ‖T‖ + ‖U‖, This shows that [N4} holds normed linear spaces and hence R or C are Banach spaces. PROVED.
In this way we have shown that B(N, N’) is a normed Linear Space. Example – 2:- The Euclidean space 𝐑𝐧 and Unitary space 𝐂 𝐧 are Normed spaces which are Banach spaces also.
STEP – III: Now We shall show that B(N, N’) is a Banach Space, whenever N’ are Banach spaces:- 1
Proof:- We define norm on Rn or C n by ‖x‖ = (∑𝑛𝑖=1|x𝑖 |2 )2 , ∀x ∈ Rn or C n . i.e. x = (x1 , x2 , ⋯ , x𝑛 )
For this we have to show that B(N, N’) is complete whenever N is complete:-
1st of all we shall verify all the four axioms of normed linear spaces:-
Let 〈Tn 〉 be a Cauchy sequence in B(N, N’)  ‖Tm − Tn ‖ < ϵ , m, n  n0 , .... (ii) 1
Now let any fixed x  N, We have ‖Tm (x) − Tn (x)‖ = ‖(Tm − Tn )𝑥‖, {∵ N ′ is a normed linear space} [N1]: By defined definition of norm on Rn or C n we have ‖x‖ = (∑𝑛𝑖=1|x𝑖 |2 )2 , ∀x ∈ Rn or C n  ‖x‖ ≥ 0, {as each |x𝑖 | ≥ 0}, Thus [N1] holds
1 1
≤ ‖Tm − Tn ‖‖𝑥‖ , {∵ ‖Tx‖ ≤ ‖T‖‖x‖ } [N2]: By defined definition of norm on Rn or C n by ‖x‖ = (∑𝑛𝑖=1|x𝑖 |2 )2 , ∀x ∈ Rn or C n we have ‖x‖ = 0,  (∑𝑛𝑖=1|x𝑖 |2 )2 = 0
Thus we get ‖Tm (x) − Tn (x)‖ ≤ ‖Tm − Tn ‖‖𝑥‖ , ... (iii)  ∑𝑛𝑖=1|x𝑖 |2 = 0  |x𝑖 | = 0, for each 𝑖  x𝑖 = 0, for each 𝑖  (x1 , x2 , ⋯ , x𝑛 ) = (0,0, ⋯ ,0)  𝑥 = 0, Thus [N2] holds
From (ii) and (iii) we get ‖Tm (x) − Tn (x)‖ < ϵ , m, n  n0 1 1
[N3]: By defined definition of norm on Rn or C n we have ‖x‖ = (∑𝑛𝑖=1|x𝑖 |2 )2 , ∀x ∈ Rn or C n , We have ‖αx‖ = (∑𝑛𝑖=1|𝛼x𝑖 |2 )2
In this way 〈Tn (x)〉 be a Cauchy sequence in N’ for each x  N. But N’ is complete and hence this sequence converges in N’. 1
  a unique y = T(x) in N’ such that 𝑙𝑖𝑚. Tn (x) = 𝑦, for a fixed x, This defines a mapping T: N → N′, s.t. T(x) = y = 𝑙𝑖𝑚. Tn (x) = |α|(∑𝑛𝑖=1|x𝑖 |2 )2 , {by propertyof mod function} This  ‖αx‖ = |α|‖x‖ , {by definition of mod function} Thus [N3] holds
1
We shall show that T is a bounded Linear Transformation:- (1) Tis a Linear Transformation:- Let any x1 , x2 ∈ N and any a, b F, we have [N4]: By defined definition of norm on Rn or C n , we have ‖x + y‖ = (∑𝑛𝑖=1|x𝑖 |2 )2 , ∀x, y ∈ Rn or C n were x = (x1 , x2 , ⋯ , x𝑛 ), y = (y1 , y2 , ⋯ , 𝑦𝑛 )
T(ax1 + bx2 ) = 𝑙𝑖𝑚T𝑛 (ax1 + bx2 ), {∵ T = 𝑙𝑖𝑚. Tn (x)} ‖x + y‖2 = ‖(x1 , x2 , ⋯ , x𝑛 ) + (y1 , y2 , ⋯ , 𝑦𝑛 )2 ‖2 = ‖(x1 + y1 , x2 + y2 , ⋯ , x𝑛 + y𝑛 )‖2 = ∑𝑛𝑖=1|x𝑖 + 𝑦𝑖 |2 , {by definition of mod}
= 𝑙𝑖𝑚{aT𝑛 (x1 ) + bT𝑛 (x2 )}, {by linearity of T𝑛 }  ‖x + y‖2 = ∑𝑛𝑖=1|x𝑖 + 𝑦𝑖 | |x𝑖 + 𝑦𝑖 | ≤ ∑𝑛𝑖=1{|x𝑖 + 𝑦𝑖 | |x𝑖 | + |x𝑖 + 𝑦𝑖 | |𝑦𝑖 |}, {by property of mod function}
= a 𝑙𝑖𝑚{T𝑛 (x1 ) + b 𝑙𝑖𝑚T𝑛 (x2 )}  ‖x + y‖2 ≤ ∑𝑛𝑖=1{|x𝑖 + 𝑦𝑖 | |x𝑖 |} + ∑𝑛𝑖=1{|x𝑖 + 𝑦𝑖 ||𝑦𝑖 |}, {by property of mod function}
= aT(x1 ) + bT(x2 ) , {∵ T = 𝑙𝑖𝑚. Tn (x)}, Thus T is a Linear Transformation.  ‖x + y‖2 ≤ ‖x + y‖‖x‖ + ‖x + y‖‖y‖, {by Cauchy inequality i. e. ∑𝑛𝑖=1|x𝑖 𝑦𝑖 | ≤ ‖x‖ ‖y‖ }
(2) T is bounded:- Let any x N, we have ‖T(x)‖ = ‖𝑙𝑖𝑚T𝑛 (x)‖, {∵ T = 𝑙𝑖𝑚. Tn (x)}  ‖x + y‖2 ≤ ‖x + y‖(‖x‖ + ‖y‖)  ‖x + y‖ ≤ ‖x‖ + ‖y‖. Thus [N4] holds.
= 𝑙𝑖𝑚‖T𝑛 (x)‖ In this way Rn or C n are normed linear spaces.
Thus we get ‖T(x)‖ ≤ 𝑙𝑖𝑚{𝑠𝑢𝑝. ‖T𝑛 ‖ ‖𝑥‖}, ... (iv) Now we shall show that Rn or C n are complete:- Let 〈x1 , x2 , ⋯ , xm ⋯ 〉 be a Cauchy sequence in Rn or C n
Now since 〈Tn 〉 be a Cauchy sequence  |Tm − Tn | < 𝜖 Since each xm is an n – tuples of real or complex numbers, we shall write x𝑚 = (x1 (𝑚) , x2 (𝑚) , ⋯ , x𝑛 (𝑚) ), so that x𝑘 (𝑚) is kth co-ordinate of x𝑚
 |‖Tm ‖ − ‖Tn ‖| ≤ ‖Tm ‖ − ‖Tn ‖ < 𝜖, {as |‖x‖ − ‖𝑦‖| ≤ ‖x‖ − ‖𝑦‖
Let 𝜖 > 0 be given also Let 〈x𝑚 〉 be a Cauchy sequence, then  a +ve integer m0 such that ∀𝑙, 𝑚 ≥ m0  ‖xm − 𝑥𝑙 ‖ < ϵ,  ‖xm − 𝑥𝑙 ‖2 < ϵ2
 ‖Tm − Tn ‖ < ϵ ,  n  n0 , ... (v) 2 2
,  ∑𝑛𝑖=1|x𝑖 (𝑚) − x𝑖 (𝑙) | < ϵ2 , ... (i)  ‖x𝑖 (𝑚) − x𝑖 (𝑙) ‖ < ϵ2 , ∀𝑖 = 1,2, … . , 𝑛 ,  ‖x𝑖 (𝑚) − x𝑖 (𝑙) ‖ < ϵ, ∀𝑖 = 1,2, … . , 𝑛
This relation shows that 〈‖Tn ‖〉 is a Cauchy sequence on real line and hence convergent and bounded so that a K > 0 s.t. sup. ‖Tn ‖ ≤ K
Thus (iv) becomes ‖T(x)‖ ≤ K‖𝑥‖, x ∈ N, In this way we have shown that T is bounded. This shows that 〈x𝑖 (𝑚) 〉∞
𝑚=1 is a Cauchy sequence of real or complex numbers for each fixed but arbitrary 𝑖

Now since T is both linear and bounded  T  B(N, N’) We know that R or C is complete,  each of these sequence converges to a point say z𝑖 in R or C, so that lim x𝑖 (𝑚) = z𝑖 , ∀𝑖 = 1,2, … , 𝑛, ... (ii)
𝑚→∞
It remains to show that 𝐓𝐧 → 𝐓 :- Since 〈Tn 〉 is a Cauchy sequence in N’  From (ii) we have ‖Tm − Tn ‖ < ϵ , m, n  n0 We now show that the Cauchy sequence 〈xm 〉 converges to the point z = (z1 , 𝑧2 , ⋯ , 𝑧𝑛 ) Rn or C n :- Let 𝑙 → ∞ then by (i) and (ii) we get
2
Let x  N such that ‖𝑥‖ ≤ 1 , Also we have Tn (x) → T(𝑥) and hence by definition of convergence we may choose a +ve integer nx ≥ n0 , such ∑𝑛𝑖=1|x𝑖 (𝑚) − z𝑖 | < ϵ2  ‖x𝑚 − z‖2 < ϵ2 ,  ‖x𝑚 − z‖ < ϵ This shows that the Cauchy sequence 〈xm 〉 converges to the point z ∈ Rn or C n
that ‖T(x) − Tnx (𝑥)‖ < 𝜖 ,  nx  n0 ... (vi) Hence Rn or C n are complete spaces consequently Rn and C n are Banach space PROVED
Now n  n0 and  ‖𝑥‖ ≤ 1 , we have ‖Tn (𝑥) − T(x)‖ = ‖Tn (𝑥) − Tnx (𝑥) + Tnx (𝑥) − T(x)‖
≤ ‖Tn (𝑥) − Tnx (𝑥)‖ + ‖Tnx (𝑥) − T(x)‖ 2017-18-6. Let B and B are Banach spaces. Let T be a linear Transformation of B into B , then T is a continuous mapping iff its graph is
= ‖(Tn − Tnx )𝑥‖ + ‖Tnx (𝑥) − T(x)‖ closed. OR State & prove closed graph theorem
≤ ‖(Tn − Tnx )‖‖𝑥‖ + ‖Tnx (𝑥) − T(x)‖ , { as ‖T(x)‖ ≤ ‖T‖‖x‖ } PROOF OF THE ONLY IF PART:- We assume that the linear Transformation from B into B, is continuous then we shall prove that Graph of T is closed
≤ ‖(Tn − Tnx )‖ + ‖Tnx (𝑥) − T(x)‖ , { as ‖𝑥‖ ≤ 1 } :- 1st we shall define Graph of T : it is denoted by T0 and defined by T0 = B × B′ = {(x, T(x)) ∶ x ∈ B, T(x) ∈ B′}
< 𝜖 + 𝜖 , { using (v) and (vi) } ̅0 = T0 :-
Proof of the Only If part:- To prove that T0 the graph of T is closed We shall show that T
In this way we have ‖Tn (𝑥) − T(x)‖ < 2𝜖,  n  n0 , ‖𝑥‖ ≤ 1 , x N ̅0 , so we need only to prove T̅0 ⊆ T0 :- Let any (x, y) ∈ T
̅0  (x, y) is an adherent point of T
̅0
By definition of closure of a set we have T0 ⊆ T
 𝑆𝑢𝑝. {‖Tn (𝑥) − T(x)‖: x N: ‖x‖ ≤ 1} < 2𝜖,  n  n0
Hence  a sequence 〈xn , T(xn )〉 in T0 such that (xn , T(xn )) → (x, y) which implies that xn → x and T(xn ) → y, Now since T is continuous
 𝑆𝑢𝑝. {‖(Tn − T)𝑥‖: x N: ‖𝑥‖ ≤ 1} < 2𝜖,  n  n0
xn → x  T(xn ) → T(x) and so y=T(x), This shows that (x, y) = (x, T(x)) ∈ T0
 ‖(Tn − T)𝑥‖ < 2𝜖,  n  n0 , {by definition of ‖T‖ }
Thus T̅0 ⊆ T0 consequently T ̅0 = T0 , Hence T0 is closed.
 𝐓𝐧 → 𝐓, Consequently B(N, N’) is a Banach Space. This completes the proof of the theorem
PROOF OF THE IF PART:- We assume that T0 is closed and then we shall show that T is continuous :- we denote B1 the sub space of B and the
norm function defined in B1 as ‖x‖1 = ‖x‖ + ‖T(x)‖
2018-5. Give two examples of Normed vector space which are Banach spaces, P – 19, 20
Now since ‖T(x)‖ ≤ ‖x‖ + ‖T(x)‖ ,  ‖T(x)‖ ≤ ‖x‖1 This shows that T is bounded and consequently continuous, Thus in order to prove the
Example – 1:- The real Liner space R and the complex linear space C are Normed spaces which are Banach spaces.also.
converse part it is sufficient to prove that B and B1 has same topology :- i.e. we shall show that B and B1 are homeomorphic:- Let the identity mapping
Proof:- We define norm on R or C by ‖x‖ = |x|, ∀x ∈ R or C.
1st of all we shall verify all the four axioms of normed linear spaces:- I: B → B1 defined by I(x) = x, xB1 obviously I is one–one and onto Also.

9 10

= ‖y
1
‖ym − yn ‖ +
| ‖yn ‖−‖ym ‖ |
≤ ‖y
1
‖ym − yn ‖ +
1
‖ym − yn ‖ = ‖x‖2 + (x, y) + (y, x) + ‖y‖2 , {by [H3] } , . . . (*)
m‖ ‖ym ‖ m‖ ‖ym ‖
1
Also we have By [H3] we have ‖x − y‖2 = (x − y, x − y) = (x, x − y) − (y, x − y), {by Linearity property of inner product in H}
Thus we get ‖xm − xn ‖ ≤ 2 ‖y ‖ym − yn ‖ ,  0,  m , n   , { by (I) } = (x, x) − (x, y) − (y, x) + (y, y), {by Linearity property of inner product in H}
m‖
yn
Hence sequence 〈xn 〉 is a Cauchy sequence in S, Since S is complete   xS, s. t. xn  x ,  → x , ....(II) = ‖x‖2 − (x, y) − (y, x) + ‖y‖2 , {by [H3] } , . . . (**)
‖yn ‖
Adding (*) & (**), we get ‖x + y‖2 + ‖x − y‖2 = 2‖x‖2 + 2‖y‖2
Also we have | ‖yn ‖ − ‖ym ‖ | ≤ ‖yn − ym ‖ ,  0,  m , n   , {by (I)}
Subtracting (**) from (*), we get ‖x + y‖2 − ‖x − y‖2 = 2(x, y) + 2(y, x) , . . . (***)
 〈‖yn ‖〉 is a Cauchy sequence of real numbers in R
Replacing y by iy in (***) we get ‖x + iy‖2 − ‖x − iy‖2 = 2(x, iy) + 2(iy, x)
Since R is complete  〈‖yn ‖〉 → α ∈ R , (say)
 ‖x + iy‖2 − ‖x − iy‖2 = 2i̅(x, y) + 2i(y, x)
 by (II) yn → αx ∈ N . Hence N is complete. This completes the proof of the theorem.
 ‖x + iy‖2 − ‖x − iy‖2 = −2i(x, y) + 2i(y, x) , . . . . (****)
2017-4. Let T be a linear transformation from a normed linear space N into another normed linear space N then Prove that the following
Multiplying by i in (****) we get  i‖x + iy‖2 − i‖x − iy‖2 = 2(x, y) − 2(y, x) , . . . . (*****)
statements are equivalent: (i)T is continuous and (ii) T is continuous at the origin.
Now (***) + (*****)  ‖x + y‖2 − ‖x − y‖2 + i‖x + iy‖2 − i‖x − iy‖2 = 2(x, y) + 2(y, x) + 2(x, y) − 2(y, x)
PROOF:- We shall Prove that (i)  (ii) and (ii)  (i)
 ‖x + y‖2 − ‖x − y‖2 + i‖x + iy‖2 − i‖x − iy‖2 = 4(x, y) PROVED.
(i)  (ii):- Given T is continuous i.e. xn → x ,  . T(xn ) → T(x) ,
2017-7(b) State and prove Pathogorian theorem for Hilbert space.
Let 𝑧𝑛 = 𝑥𝑛 − 𝑥 be sequence of vectors in N and 𝑧𝑛 → 0 as 𝑥𝑛 → 𝑥, Now 𝑇(𝑧𝑛 ) = 𝑇(𝑥𝑛 − 𝑥) = 𝑇(𝑥𝑛 ) − 𝑇(𝑥) → 0, {by assumption}
Thus we see that 𝑧𝑛 → 0  𝑇(𝑧𝑛 ) → 0,  T is continuous at origin and thus (i)  (ii): established
Now we shall show that (ii)  (i):- Given T is continuous at origin i.e. xn → 0 ,  . T(xn ) → 0, Let xn → x , so that 𝑦 = xn − x → 0 ,
𝑇(𝑦) → 0,  𝑇(xn − x) → 0, {as T is continuous at origin} 2017-8. State and prove Reistz’s representation theorem
 𝑇(xn ) − 𝑇(x) → 0,  𝑇(xn ) → 𝑇(x),  T is continuous and thus (ii)  (i): established. This completes the proof of the theorem. ANSWER:- SEE Q. No. 8 of MSC 10 -2018
2017-5. State and prove closed graph theorem 2017-9(a) Define normal operator on a Hilbert space H, If 𝐍𝟏 & 𝐍𝟐 are normal operator on H, with property that either commutes with the
ANSWER:- SEE Q. No. 6 of MSC 10 -2018 adjoint of the other then 𝐍𝟏 + 𝐍𝟐 and 𝐍𝟏 ∙ 𝐍𝟐 are normal P-375
2017-6(b) Give an example of a Banach space which is not a Hilbert space. PROOF:- Since N1 & N2 are normal operator on H  N1 N1 ∗ = N1 ∗ N1 Also N2 N2 ∗ = N2 ∗ N2 ,….(i)
The space C[0,1] of continuous functions f:[0,1]→R with the supremum norm is an example of a Banach space Also it is given that N1 N2 ∗ = N2 ∗ N1 and N2 N1 ∗ = N1 ∗ N2 ,….(ii)
which is not a Hilbert space. We need to check that the parallelogram law is not satisfied. Take f(x)=x, x∈[0,1], 1st we shall show that N1 + N2 is a normal operator on H:- We have (N1 + N2 )(N1 + N2 )∗ = (N1 + N2 )(N1 ∗ + N2 ∗ )
and g(x)=1, x∈[0,1] = N1 N1 ∗ + N1 N2 ∗ + N2 N1 ∗ + N2 N2 ∗ = N1 ∗ N1 + N2 ∗ N1 + N1 ∗ N2 + N2 ∗ N2 , {on using (i) & (ii)}
= N1 ∗ (N1 + N2 ) + N2 ∗ (N1 + N2 ) = (N1 ∗ + N2 ∗ )(N1 + N2 ) = (N1 + N2 )∗ (N1 + N2 )
2017-6. (a) Give an example of an inner product space which is not a Hilbert space. ∗
In this way we get (N1 + N2 )(N1 + N2 ) = (N1 + N2 )∗ (N1 + N2 ) This shows that N1 + N2 is normal
An inner product space inner product space is a vector space with an inner product defined on it. An inner product Now we shall show that N1 N2 is a normal operator on H:-
on X defines a norm on X given by ∥x∥=⟨x, x⟩−−−−−√‖x‖=⟨x, x⟩ We have (N1 N2 )(N1 N2 )∗ = (N1 N2 )(N2 ∗ + N1 ∗ ) = N1 (N2 N2 ∗ )N1 ∗ , {on using associativity of operators}
= N1 (N2 ∗ N2 )N1 ∗ , , {on using (i) }
and a metric on X given by d(x,y)=∥x−y∥=⟨x−y,x−y⟩.
= (N1 N2 ∗ )(N2 N1 ∗ ) , {on using associativity of operators}
A complete inner product space is called a Hilbert spaceHilbert space. Hence inner product spaces are normed = (N2 ∗ N1 )(N1 ∗ N2 ) , {on using (ii) }
linear spaces and Hilbert spaces are Banach spaces. = N2 ∗ (N1 N1 ∗ )N2 , {on using associativity of operators}
A simple straightforward calculation shows that a norm on an inner product space satisfies the = N2 ∗ (N1 ∗ N1 )N2 , {on using (i) }
important parallelogram law ∥x+y∥2+∥x−y∥2=2(∥x∥2+∥y∥2). = (N2 ∗ N1 ∗ )(N1 N2 ), {on using associativity of operators}
We conclude that if a norm does not satisfy the parallelogram law, it cannot be obtained from an inner = (N1 N2 )∗ (N1 N2 ), {on using property of normal operators}
In this way we get (N1 N2 )(N1 N2 )∗ = (N1 N2 )∗ (N1 N2 ) This shows that N1 N2 is normal operator; This completes the proof of the theorm.
product. Without risking misunderstanding we may thus say:
Not all normed linear spaces are inner product spaces. Not all normed linear spaces are inner product spaces. 2017-9(b) An operator T on a Hilbert space H is unitary iff it is an isometric isomorphism of H onto itself.
The space Lp with p≠2 is not an inner product space, hence not a Hilbert space. ANSWER:- SEE Q. No. 9(a) of MSC 10 -2016
Our statement means that the norm of lp with p≠2 cannot be obtained from an inner product. We prove this by showing ________________________________________________________________________________________________________________________
that the norm does not satisfy the parallelogram law. In fact, let us take x =(1,1,0,0,…) ∈ lp and y=(1,−1,0,0,…)∈lp 2014-16-2. Define a Banach space and prove that the set 𝐑𝐧 of all n tuples 𝐱 = (𝐱 𝟏 , 𝐱 𝟐 , ⋯ , 𝐱 𝐧 ) of real numbers is a real Banach Space if for
𝐱 = (𝐱 𝟏 , 𝐱 𝟐 , ⋯ , 𝐱 𝐧 )𝐑𝐧 and 𝐲 = (𝐲𝟏 , 𝐲𝟐 , ⋯ , 𝐲𝐧 )𝐑𝐧 and R we define (𝐱 𝟏 , 𝐱 𝟐 , ⋯ , 𝐱 𝐧 ) + (𝐲𝟏 , 𝐲𝟐 , ⋯ , 𝐲𝐧 ) =
Then ∥x∥=∥y∥=21p,∥x+y∥=∥x−y∥=2.‖x‖=‖y‖=21p,‖x+y‖=‖x−y‖=2.
(𝐱 𝟏 + 𝐲𝟏 , 𝐱 𝟐 + 𝐲𝟐 , ⋯ , 𝐱 𝐧 + 𝐲𝐧 ), (𝐱 𝟏 , 𝐱 𝟐 , ⋯ , 𝐱 𝐧 ) = (𝐱 𝟏 , 𝐱𝟐 , ⋯ , 𝐱𝐧 ) and ‖𝐱‖ = (∑𝐧𝐢=𝟏(𝐱 𝐢 )𝟐 )𝟏/𝟐 as norm of x
We now see that parallelogram law is not satisfied if p≠2. We also know that lp is complete. ANSWER:- Definition of a Banach space: A normed linear space (B, ‖ ‖) which is complete as a metric space is called a Banach Space. Or in other
Hence lp with p≠2 is a Banach space which is not a Hilbert space. words the normed linear space (B, ‖ ‖) is said to be a Banach Space if every Cauchy sequence in B converges in B.
2017-6. (a) Give an example of an inner product space which is not a Hilbert space. 1st we shall show that Rn is a normed linear space under the given norm function:- Verification of axioms of Normed Linear space:-
Any non-closed subspace of a Hilbert space will do. For example, the linear span of a Hilbert basis of an infinite- [N1]:- By given definition of norm function we have ∀ x ∈ Rn , where x = (x1 , x2 , x3 , ⋯ , xn ),
‖x‖ = (∑ni=1|xi |2 )1/2 , Now since |xi | ≥ 0,  i= 1, 2, 3,…, n,  (∑ni=1|xi |2 )1/2 ≥ 0,  ‖x‖ ≥ 0, Thus [N1] holds in Rn
dimensional Hilbert space.
[N2]:- Let any x, y ∈ Rn , where x = (x1 , x2 , x3 , ⋯ , xn ), y = (y1 , y2 , y3 , ⋯ , yn ) and xi , yi R
For an explicit example, take H = ℓ2(R), the vector space of square summable sequences of real numbers with Let ‖x‖ = 𝟎,  (∑ni=1|xi |2 )1/2 = 0 , {By given definition of norm function}
pointwise operations and inner product ⟨{an},{bn}⟩= ∑∞ k=1 a k bk .  ∑ni=1|xi |2 = 0 , {on squaring both sides}
Let ei be the sequence that has a1 in the ith coordinate and 0s elsewhere. Then {ei∣i∈N}is a Hilbert basis  xi = 0, {as: sum of squares of real numbers is 0 this is possible iff each xi = 0}
for H, and its linear span (the vector space of all almost null sequences) is an inner product space that is not  x1 , x2 , x3 , ⋯ , xn = 0
complete.  (x1 , x2 , x3 , ⋯ , xn ) = 0
 x = 0, Thus [N2] holds in Rn
2017-7(a) Establish polarization identity in a Hilbert space.
[N3]:- Let any x, y ∈ Rn , where x = (x1 , x2 , x3 , ⋯ , xn ), y = (y1 , y2 , y3 , ⋯ , yn ) and xi , yi R
ANSWER:- Let H be a Hilbert space , Let x, y ∈ H then we shall show that 𝟒(𝐱, 𝐲) = ‖𝐱 + 𝐲‖𝟐 − ‖𝐱 − 𝐲‖𝟐 + 𝐢‖𝐱 + 𝐢𝐲‖𝟐 − 𝐢‖𝐱 − 𝐢𝐲‖𝟐 ,
We have ‖x + y‖2 = ‖(x1 , x2 , x3 , ⋯ , xn ) + (y1 , y2 , y3 , ⋯ , yn )‖2
PROOF:- By [H3] we have ‖x + y‖2 = (x + y, x + y) = (x, x + y) + (y, x + y), {by Linearity property of inner product in H}
= ‖(x1 + y1 , x2 + y2 , x3 + y3 , ⋯ , xn + y𝑛 )‖2 ,{by defined definition of addition inRn }
= (x, x) + (x, y) + (y, x) + (y, y), {by Linearity property of inner product in H}
 11 12

= ∑ni=1|xi + 𝑦𝑖 |2 , {By given definition of norm function} 2016-4. A linear transformation T from a Normed linear space E into a Normed linear space F is continuous iff T is bounded in the sense that
≤ ∑ni=1|xi + 𝑦𝑖 |(|xi | + |𝑦𝑖 |) , {on using property of mod function}  a +ve real number M, s. t. ‖𝐓𝐱‖ ≤ 𝐌‖𝐱‖ ,  x  E.
= ∑ni=1|xi + 𝑦𝑖 ||xi | + ∑ni=1|xi + 𝑦𝑖 ||𝑦𝑖 |, {as summation distributed term wise} PROOF:- 1ST PART:- Let T be a linear transformation from a Normed linear space E into a Normed linear space. We assume that T is continuous then
≤ ‖x + y‖‖x‖ + ‖x + y‖‖y‖, {by Cauchy’s inequality} we shall prove that T is bounded:- Given T is continuous i.e. xn → x ,  . T(xn ) → T(x), In order to prove that T is bounded we have to Prove that
Thus we get ‖x + y‖2 ≤ ‖x + y‖(‖x‖ + ‖y‖) ,  x  E, where M is a +ve real number :- If possible let There does not exist such M satisfying this condition ‖Tx‖ ≤ M‖x‖, Hence for each +ve
 ‖x + y‖ ≤ ‖x‖ + ‖y‖, Thus [N3] holds in Rn integer n we can find a vector xn such that ‖T(xn )‖ ≥ 𝑛‖xn ‖ 
‖T(xn )‖
≥ 1 , Now since 𝑛‖xn ‖ is a +ve real scalar  This can be written as
𝑛‖xn ‖
[N4]:- Let any x ∈ Rn , any R, where x = (x1 , x2 , x3 , ⋯ , xn ), now by definition of scalar multiplication in Rn , xn xn xn 1 1
we have ‖x‖ = ‖ (x1 , x2 , x3 , ⋯ , xn )‖ = ‖x1 , x2 , x3 , ⋯ , xn ‖ ‖T ( ‖
)‖ ≥ 1  ‖T(y𝑛 )‖ ≥ 1 , where y𝑛 =  ‖y𝑛 ‖ = ‖ ‖= ‖xn ‖ = → 0 as n → ∞, Thus we get ‖y𝑛 ‖ → 0
𝑛‖xn 𝑛‖xn ‖ 𝑛‖xn ‖ 𝑛 ‖xn ‖ 𝑛
= (∑ni=1|xi |2 )1/2 , {By given definition of norm function} But ‖T(y𝑛 )‖ ≥ 1 i.e. ‖T(y𝑛 )‖ does not tends to 0. This contradicts the assumption that T is continuous. Hence our assumption that T is not
= ||(∑ni=1|xi |2 )1/2 bounded was wrong, Consequently T is bounded. i.e. ‖Tx‖ ≤ M‖x‖ ,  x  E. This proves 1st part
= || ‖x‖ , {By given definition of norm function}, In this way Rn is a normed linear space under the given norm function. Converse Part:- Now we assume that T is bounded i.e. ‖Tx‖ ≤ M‖x‖ ,  x  E.and then we shall show that T is continuous:- Let xn → x
Now we shall show that 𝐑𝐧 is complete as a metric space:-Let 〈x1 , x2 , x3 , ⋯ , xm , ⋯ 〉 be a Cauchy sequence in Rn , since each xm is n-tuples  ‖xn ‖ → ‖x‖, {as norm is a continuous function} ... (*)
of real numbers, So we shall write xm = (x1 (𝑚) , x2 (𝑚) , x3 (𝑚) , ⋯ , xn (𝑚) )  ‖T(xn )‖ ≤ M‖x𝑛 ‖ , {by assumption of converse part}
Now since 〈xm 〉 is a Cauchy sequence,   a +ve real number m0 , s. t.  𝑙, 𝑚 ≥ m0  ‖xl − xm ‖ < ϵ where ϵ is arbitrary small +ve real number  ‖T(xn )‖ → M‖𝑥‖ , {by (*)},  T is continuous. This completes the proof of the theorem.
 ‖x𝑙 − xm ‖2 < ϵ2 , {on squaring both sides} 2016-5. State and prove the open mapping theorem. ML-85-91, [JN-295 – 299]
2
 ∑ni=1|xi (𝑙) − 𝑥𝑖 (𝑚) | < ϵ2 , … (i), {By given definition of norm function} STATEMENT:- Let B and B’ are Banach spaces. Let T be a continuous linear Transformation of B onto B , then T is open mapping.
 ‖x𝑖 (𝑙) − x𝑖 (𝑚) ‖ < ϵ , This shows that the sequence 〈x𝑖 (𝑚) 〉 is a Cauchy sequence in Rn . Proof is very lengthy about 5 pages.
Now since R is complete , each of these sequence 〈x𝑖 (𝑚) 〉 converges to point z𝑖 (say) 2016-6. Define inner product space and a Hilbert space. Prove that an inner product space is a Normed linear space with respect to the norm
Thus we get lim {x𝑖 (𝑚) } = z𝑖 ,  i=1, 2, 3, …..,n defined by ‖𝐱‖ = +√𝐱|𝐱 i.e. ‖𝐱‖ = +√(𝒙, 𝐱) P – 119
𝑚→∞
2
Now taking 𝑙 → ∞ in (i) we get ∑ni=1|xi (𝑚) − 𝑧𝑖 | < ϵ2 PROOF:- Let V(F) be an inner product space then we shall show that V(F) is a Normed space w.r.t. the norm defined by ∀x ∈ V  ‖x‖ = +√(𝑥, x)
 ‖𝑥𝑚 − z‖2 < ϵ2 , {By given definition of norm function} or ‖x‖2 = (x, x)
 ‖𝑥𝑚 − z‖ < ϵ ,  Cauchy sequence 〈xm 〉 converges to z  Rn Now we shall verify all the four axioms of a normed space w.r.t this norm function :-
Hence Rn is a complete metric space, consequently Rn is a Banach space. PROVED [N1]:- By defined definition of norm function we have ‖x‖2 = (x, x) ≥ 0, {by 1st axiom of I.P.V.S.} Thus [N1] holds
2016-3. Let M be a closed linear subspace of a Normed linear space N if the norm of the coset [x] = x+M in the quotient space N/M is defined [N2]:- By defined definition of norm function we have ‖x‖2 = (x, x), Assuming that ‖x‖ = 0,
by ‖𝐱 + 𝐌‖ = 𝐢𝐧𝐟{‖𝐱 + 𝐯‖: 𝐯 ∈ 𝐌} then prove that N/M is a Normed linear space.  ‖x‖2 = 0, (x, x) = 0,  x = 0, {by defn. of I.P.}.Thus [N2] holds
PROOF: Let M be a closed linear subspace of a normal linear space N, Let the norm of a coset x +M in the quotient space N/M defined by [N3]:- By defined definition of norm function we have ‖x‖2 = (x, x),  ‖αx‖2 = (αx, αx) = 𝛼𝛼̅(x, x) , {by definition. of I.P.}
‖x + M‖ = inf. {‖x + m‖: m ∈ M} then 1st we shall show that N/M is a normed linear space: we shall verify all the postulates of norm function for = |𝛼|2 ‖x‖2 Thus we get ‖αx‖ = |𝛼|‖x‖, Thus [N3] holds
the defined normed function on quotient space N/M= {x + M: x ∈ N} [N4]:- By defined definition of norm function we have ‖x + y‖2 = (x + y, x + y) = (x, x) + (x, y) + (𝑦, x) + (y, y), {by definition. of I.P.}
[N1]: By given definition of norm function defined in N/M we have ‖x + M‖ = inf. {‖x + m‖: m ∈ M} = ‖x‖2 + (x, y) + ̅̅̅̅̅̅̅
(x, y) + ‖𝑦‖2 , {by definition. of I.P.}
Now since ‖x + m‖ is a non –ve real number and every set of non –ve real numbers is bounded below. it follows that inf. {‖x + m‖: m ∈ M} exist = ‖x‖2 + ‖𝑦‖2 + 2𝑅𝑒(x, y) , {by definition. of I.P.}
and is non –ve real number also. 2 2
≤ ‖x‖ + ‖𝑦‖ + 2|(x, y)| , {by property of mod function}
‖x + M‖ ≥ 0,  x + MN/M ≤ ‖x‖2 + ‖𝑦‖2 + 2‖𝑥‖‖y‖ , {by Cauchy Schwarz’s in equality}
Thus [N1] holds for the given function to be a norm function in N/M = (‖𝑥‖ + ‖y‖)2
[N2]: in the quotient space N/M the zero element is M, Let x + M ∈ N/M, be the any element of N/M, Let x + M = M, 2 2
Thus we get ‖x + y‖ ≤ (‖𝑥‖ + ‖y‖) , ‖x + y‖ ≤ ‖𝑥‖ + ‖y‖, Thus [N4] holds
 x ∈ M, {by property of cosets}, Hence ‖x + M‖ = inf. {‖x + m‖: m ∈ M, x ∈ M} In this way the inner product space is a Normed linear space. This completes the proof of the theorem.
 ‖x + M‖ = inf. {‖y‖: y ∈ M} = 0,{as M being a subspace containing a zero vector whose norm is 0}
Thus x + M = M,  ‖x + M‖ = 0 2016-7. State and prove projection theorem in a Hilbert space completely.
conversely Let ‖x + M‖ = 0, then we shall show that Let x + M = M :- Statement: - “If M is a closed subspace of a Hilbert space H then: H = M ⨁ M ⊥ " To prove the theorem we have to prove that: (i) M ⊥ is a closed
Since ‖x + M‖ = 0,  inf. {‖x + m‖: m ∈ M} = 0 subspace of H, (ii) M + M ⊥ is a closed subspace of H, (iii) M ∩ M ⊥ = {𝟎} , (iv) H = M + M ⊥ ,
(i) If M be a non-empty subset of a Hilbert space H, then M ⊥ is a closed linear subspace of H
  a sequence 〈𝑚k 〉 ∈ M, such that ‖x + m𝑘 ‖ → 0, as k ,  lim mk → −x
k→∞ PROOF:- 1st we shall show that 𝐌 ⊥ is a linear subspace of H:- For this we shall show that for any x1 , x2 ∈ M ⊥ and for any two scalars  ,  
But 〈𝑚k 〉 ∈ M, this sequence is closed so that lim mk belongs to M and hence −x ∈ M
k→∞ αx1 + βx2 ∈ M ⊥ i.e. (αx1 + βx2 , y) = 0, y ∈ M:- Let any x1 , x2 ∈ M ⊥  (x1 , y) = 0, (x2 , y) = 0, y ∈ M } , . . . (i)
and consequently x ∈ M  x + M = M, {by Prop. of cosets} Now yM, we have (αx1 + βx2 , y) = α(x1 , y) + β(x2 , y),{by linearity property of inner product}
Thus [N2] holds for the given function to be a norm function in N/M  (αx1 + βx2 , y) = α0 + β0 = 0 + 0 = 0 {Using (i) & property of liner space}
[N3]: in the quotient space N/M we have to show that ‖α(x + M)‖ = |α|‖(x + M)‖ :-  αx1 + βx2 ∈ M ⊥ , Hence that M ⊥ is a linear subspace of H.
We have ‖α(x + M)‖ = inf. {‖𝛼(x + m)‖: 𝑚 ∈ 𝑀} = inf. {|α|‖(x + m)‖: 𝑚 ∈ 𝑀} = |α| inf. {‖(x + m)‖: 𝑚 ∈ 𝑀} = |α|‖(x + M)‖ Now we shall show that 𝐌 ⊥ is a closed subspace of H:- For this we shall show that S ⊥ , contains its all limit points :- Let x be the any limit point of S ⊥ ,
Thus [N3] holds for the given function to be a norm function in N/M   a sequence 〈xn 〉 in S ⊥ , such that xn → x, Now Let yS and xn ∈ S ⊥ (xn , y) = 0 ,  n,  lim {(xn , y)} = 0,
n→∞
[N4]: in the quotient space N/M we have to show that ‖(x + M) + (y + M)‖ ≤ ‖x + M‖ + ‖𝑦 + 𝑀‖ :-
 lim (xn , y) = 0 , {as inner product is continuous mapping}
We have ‖(x + M) + (y + M)‖ = ‖(x + y) + M‖, { by definition of addition of two cosets} n→∞
= inf. {‖x + y + m‖: 𝑚 ∈ 𝑀}, { by definition of addition of norm in N/M}  (x, y) = 0 , { as lim {xn } = x}  x  M ⊥ , Hence M ⊥ is a closed subspace of H.
n→∞
= inf. {‖x + y + m1 + m2 ‖: 𝑚 = m1 + m2 ∈ 𝑀} (ii): 𝐌 + 𝐌 ⊥ is a closed subspace of H :- Now we shall show that For this we shall show that M + M ⊥ contains its all limit points:-
= inf. {‖(x + m1 ) + (𝑦 + m2 )‖: m1 , m2 ∈ 𝑀} Let z be the any limit point of M + M ⊥ ,   a sequence 〈zn 〉 in M + M ⊥ such that zn → z
≤ inf. {‖(x + m1 )‖ ∶ m1 ∈ 𝑀} + inf. {‖(x + m2 )‖ ∶ m2 ∈ 𝑀}, { by [N4] in N} Now Since M ∩ M ⊥ = {𝟎} ,  M + M ⊥ is direct sum of the subspaces M and M ⊥ consequently each zn can be uniquely written as : zn = xn + yn
≤ ‖x + M‖ + ‖𝑦 + 𝑀‖, {by definition of norm in N/M} , where xn ∈ M , yn ∈ M ⊥ , Considering the two vectors zm , zn ∈ 〈zn 〉, where zm = xm + ym and zn = xn + yn
Hence we get ‖(x + M) + (y + M)‖ ≤ ‖x + M‖ + ‖𝑦 + 𝑀‖, Thus [N4] holds for the given function to be a norm function in N/M Now we have ‖zm − zn ‖ = ‖(xm + ym ) − (xn + yn )‖ = ‖(xm − xn ) − (ym − yn )‖ , . . . . (*)
In this way N/M is a Normed linear space. This completes the proof of the theorem. Now since xm − xn ∈ M and ym − yn ∈ M ⊥  (xm − xn ) ⊥ (ym − yn ) ,  by Pythagorean theorem we have
‖(xm − xn ) − (ym − yn )‖2 = ‖xm − xn ‖2 + ‖ym − yn ‖2

13 14

Thus by (*) relation we get ‖zm − zn ‖2 = ‖xm − xn ‖2 + ‖ym − yn ‖2 , . . . . (**) = (x, Ty), { by property of inner product }
Now since 〈zn 〉 is a Cauchy sequence in Hilbert Space H  m , n → ∝  ‖zm − zn ‖ → 0 Thus we get (x, T ∗∗ y) = (x, Ty)  T ∗∗ = T, {by uniqueness of adjoint operators}. This proves (d)
And thus by (**) result we get ‖xm − xn ‖ → 0, ‖ym − yn ‖ → 0,  m, n → ∝ (i) ‖𝐓 ∗ ‖ = ‖𝐓‖: PROOF:- for any vector y in H, we have ‖T ∗ y‖2 = (T ∗ y, T ∗ y), {by property of inner product}
Consequently 〈xn 〉 is a Cauchy sequence in M and 〈yn 〉 is a Cauchy sequence in M ⊥ = (TT ∗ y, y), {by property of adjoint operator}
We know that every closed subspace of a complete metric space is complete,  H is complete and M, M ⊥ are closed subspaces of H, = |(TT ∗ y, y)|, {as (T ∗ y, T ∗ y) = ‖T ∗ y‖2 is a +ve real number}
 M, M ⊥ are complete, The Cauchy sequences〈xn 〉, 〈yn 〉 in M and M ⊥ respectively are convergent sequences in M and M ⊥ ≤ ‖TT ∗ 𝑦‖ ‖y‖ ,{By Schwarz inequality: |(x, y)| ≤ ‖x‖‖y‖ }
  two vectors x , y in M and M ⊥ respectively such that xn → x, yn → y,  m, n → ∝ , ≤ ‖T‖ ‖T ∗ 𝑦‖ ‖y‖ , {as ‖Tx‖ ≤ ‖T‖ ‖𝑥‖ }
Now z = lim zn = lim {xn + yn } = lim {xn } + lim {yn } = x + y ∈ M + M ⊥ Thus we get ‖T ∗ y‖2 ≤ ‖T‖ ‖T ∗ 𝑦‖ ‖y‖, yH
n→∝ n→∝ n→∝ n→∝
 ‖T ∗ y‖ ≤ ‖T‖‖y‖, yH … (*)
Thus z is a limit point of M + M ⊥ consequently M + M ⊥ is a closed subspace of H.
Now we have ‖T ∗ ‖ = sup. {‖T ∗ y‖: ‖y‖ ≤ 1}
(iii) 𝐌 ∩ 𝐌 ⊥ = {𝟎}:- Let any x  M ∩ M ⊥ ,  x  M and x M ⊥ ,  M ⊥ M ⊥  (x, x) = 0,  ‖x‖ = 0, {by definition of inner product}
From (*) we see that when ‖y‖ ≤ 1 then ‖T ∗ y‖ ≤ ‖T‖
 x = 0, {by [N2] }
 sup. {‖T ∗ y‖: ‖y‖ ≤ 1} ≤ ‖T‖  ‖T ∗ ‖ ≤ ‖T‖, … (**)
(iv) 𝐇 = 𝐌 + 𝐌 ⊥ :- Since M + M ⊥ is a closed subspace of H Putting N = M + M ⊥ , . . . (***)
Now applying the result (**) for the operator T* in place of operator T, we get ‖(T ∗ )∗ ‖ ≤ ‖T ∗ ‖
Then we shall show that N = H :- From (***) we have M  N , M ⊥  N  by the property of orthogonal space we have N ⊥  M ⊥ , N ⊥  M ⊥⊥
 ‖T ∗∗ ‖ ≤ ‖T ∗ ‖  ‖T‖ ≤ ‖T ∗ ‖, … (***), {as T ∗∗ = T }
 N ⊥  M ⊥ ∩ M ⊥⊥ ,  N ⊥  {0}, { ∵ M ∩ M ⊥ = {𝟎} }
Combining (**) and (***), we get ‖T‖ = ‖T ∗ ‖
 N ⊥ = {0}, { ∵ {𝟎}  N ⊥ }
(ii) ‖𝐓 ∗ 𝐓‖ = ‖𝐓‖𝟐
 (N ⊥ )⊥ = {0}⊥ , { on taking orthogonality both sides }
PROOF:- We have ‖T ∗ T‖ ≤ ‖T ∗ ‖‖T‖, {as ‖ST‖ ≤ ‖S‖‖T‖}
 N = H, { ∵ {0}⊥ = H and (N ⊥ )⊥ = N }
= ‖T‖‖T‖ ≤ ‖T‖2 , {as ‖T‖ = ‖T ∗ ‖ }
Thus we get H = M + M ⊥ , Consequently H = M ⨁ M ⊥ . This completes the proof of projection theorem.
Thus we get ‖T ∗ T‖ ≤ ‖T‖2 , … (*)
2016-4. A linear transformation T from a Normed linear space E into a Normed linear space F is continuous iff T is bounded in the sense that
further for every x  H, we have ‖Tx‖2 = (Tx, Tx), {by property of inner product}
 a +ve real number M, s. t. ‖𝐓𝐱‖ ≤ 𝐌‖𝐱‖ ,  x  E.
PROOF:- 1ST PART:- Let T be a linear transformation from a Normed linear space E into a Normed linear space. We assume that T is continuous then = (T ∗ Tx, x), {by property of adjoint operator}
we shall prove that T is bounded:- Given T is continuous i.e. xn → x ,  . T(xn ) → T(x), In order to prove that T is bounded we have to Prove that = |(T ∗ Tx, x)|, {as(T ∗ Tx, x) = ‖Tx‖2 is a +ve real number}
,  x  E, where M is a +ve real number :- If possible let There does not exist such M satisfying this condition ‖Tx‖ ≤ M‖x‖, Hence for each +ve ≤ ‖T ∗ Tx‖ ‖x‖ ,{By Schwarz inequality: |(x, y)| ≤ ‖x‖‖y‖ }
integer n we can find a vector xn such that ‖T(xn )‖ ≥ 𝑛‖xn ‖ 
‖T(xn )‖
≥ 1 , Now since 𝑛‖xn ‖ is a +ve real scalar  This can be written as ≤ ‖T ∗ 𝑇‖ ‖x‖‖x‖ , {as ‖Tx‖ ≤ ‖T‖ ‖𝑥‖ }
𝑛‖xn ‖
xn xn xn 1 1 Thus we get ‖Tx‖2 ≤ ‖T ∗ 𝑇‖ ‖x‖2 ,  xH, …. (**)
‖T ( )‖ ≥ 1  ‖T(y𝑛 )‖ ≥ 1 , where y𝑛 =  ‖y𝑛 ‖ = ‖ ‖= ‖xn ‖ = → 0 as n → ∞, Thus we get ‖y𝑛 ‖ → 0
𝑛‖xn ‖ 𝑛‖xn ‖ 𝑛‖xn ‖ 𝑛 ‖xn ‖ 𝑛 Now we have ‖𝑇‖ = sup. {‖Tx‖: ‖x‖ ≤ 1}
But ‖T(y𝑛 )‖ ≥ 1 i.e. ‖T(y𝑛 )‖ does not tends to 0. This contradicts the assumption that T is continuous. Hence our assumption that T is not  ‖𝑇‖2 = [sup. {‖Tx‖: ‖x‖ ≤ 1}]2
bounded was wrong, Consequently T is bounded. i.e. ‖Tx‖ ≤ M‖x‖ ,  x  E. This proves 1st part
 ‖𝑇‖2 = sup. {‖Tx‖2 : ‖x‖ ≤ 1}
Converse Part:- Now we assume that T is bounded i.e. ‖Tx‖ ≤ M‖x‖ ,  x  E.and then we shall show that T is continuous:- Let xn → x
From (**) we see that if ‖x‖ ≤ 1 then ‖Tx‖2 ≤ ‖T ∗ 𝑇‖
 ‖xn ‖ → ‖x‖, {as norm is a continuous function} ... (*)
 ‖T(xn )‖ ≤ M‖x𝑛 ‖ , {by assumption of converse part}  sup. {‖Tx‖2 : ‖x‖ ≤ 1} ≤ ‖T ∗ T‖,  ‖T‖2 ≤ ‖T ∗ T‖, …. (***)
 ‖T(xn )‖ → M‖𝑥‖ , {by (*)}  T is continuous. This completes the proof of the theorem. Combining (*) and (***) we get ‖T ∗ T‖ = ‖T‖2 , THIS PROVES (ii)
2016-8. Prove that for an adjoint operator T T* on B(H) the following results hold: (a) (𝐓𝟏 + 𝐓𝟐 )∗ = 𝐓𝟏 ∗ + 𝐓𝟐 ∗ , (b) (𝛂𝐓)∗ = 𝛂 ̅ 𝐓∗ , 2016-9(a) Define unitary operator on a Hilbert space H, Prove that an operator T on a Hilbert space H is unitary iff it is an isometric
(c) (𝐓𝟏 𝐓𝟐 )∗ = 𝐓𝟐 ∗ 𝐓𝟏 ∗ , (d) (𝐓 ∗ )∗ = 𝐓 isomorphism of H onto itself.
PROOF:- Let T be an adjoint operator T T* on B(H), where B(H) is a complex Banach space on the Hilbert space H. PROOF:- 1ST PART:- Assuming that T is unitary operator on a Hilbert Space H then we shall show that T is an isometric isomorphism of H onto itself:-
Proof of (a) :- Let any x, y  H, We have (𝑥, (T1 + T2 )∗ 𝑦) = ((T1 + T2 )𝑥, 𝑦), {by definition of adjoint operator (T1 + T2 ) } Since T is unitary operator on H then T is invertible, { as TT ∗ = T ∗ T = 𝐼 }
= (T1 𝑥 + T2 𝑥, 𝑦), {by linearity property of operators}  T is onto. Also since T ∗ T = 𝐼,  ‖Tx‖ = ‖𝑥‖, xH, Thus T preserves norms and so T is an isometric isomorphism of H onto itself
= (T1 𝑥, 𝑦) + (T2 𝑥, 𝑦), {by linearity property of adjoint operators} Converse Part:- We assume that T is an isometric isomorphism of H onto itself, then we shall show that T is unitary operator on H:- Since T is an
= (𝑥, T1 ∗ 𝑦) + (𝑥, T2 ∗ 𝑦), {by definition of adjoint operators T1 & T2 } isometric isomorphism of H onto itself, then T is one-one and onto  T-1 exists, Also T is isometric isomorphism  ‖Tx‖ = ‖𝑥‖, xH,
= (𝑥, T1 ∗ 𝑦 + T2 ∗ 𝑦), {by linearity property of adjoint operators}  T ∗ T = 𝐼,  (T ∗ T)T −1 = 𝐼T −1 , {post operating by T −1 }
= (𝑥, (T1 ∗ 𝑦 + T2 ∗ )𝑦), {by linearity property of operators}  T ∗ (TT −1 ) = T −1 , { by associavity of Linear operator}
= (𝑥, (T1 ∗ + T2 ∗ )𝑦), {by linearity property of operators  T ∗ (I) = T −1 , { by property of Linear operator, i.e. TT −1 = 𝐼}
Thus we get (𝑥, (T1 + T2 )∗ 𝑦) = (𝑥, (T1 ∗ + T2 ∗ )𝑦)  (T1 + T2 )∗ = T1 ∗ + T2 ∗ , {by uniqueness of adjoint operator}. This proves (a)  T ∗ = T −1 ,  𝑇T ∗ = TT −1 , {pre-operating by T}
Proof of (b):- Let any x, y  H, We have (x, (T)∗ y) = ((T)x, y), {by definition of adjoint operator (T) }  𝑇T ∗ = 𝐼 {as TT −1 = 𝐼 }, Thus we get TT ∗ = 𝐼 = T ∗ T  T is unitary operator on H. This completes the proof of the theorem.
= ((Tx), y) = (Tx, y) = (x, T ∗ y), {by definition of adjoint operator T } 2016-9(b) An operator T on a Hilbert space H is a normal operator iff its real and imaginary parts commute.
= (x, 
̅ (T ∗ y)), {by property of inner product } PROOF:-1ST PART:- Assuming that T is a normal operator on a Hilbert Space H then we shall show that the real part T 1 and the imaginary part T2 of T
= (x, (
̅ T ∗ )𝑦), , {by property of operator } commutes; i.e. T1 T2 = T2 T1 :- Since T = T1 + 𝑖T2 and T is a normal operator  T1 & T2 are self adjoint operators, i.e. T1 = T1 ∗ & T2 = T2 ∗ ,
Now we have T ∗ = (T1 + 𝑖T2 )∗ = T1 ∗ + (𝑖T2 )∗ = T1 ∗ + 𝑖̅ T2 ∗ = T1 − 𝑖 T2 , Thus we get T ∗ = T1 − 𝑖 T2 ,
Thus we get (x, (T)∗ y) = (x, ( ̅ T ∗ )𝑦)  (T)∗ = ̅ T ∗ , {by uniqueness of adjoint operators }. This proves (b)
Now we have TT ∗ = (T1 + 𝑖 T2 )(T1 − 𝑖 T2 ) = (T1 2 + T2 2 ) + 𝑖( T2 T1 − T1 T2 ), … (i)
Proof of (c) :- Let any x, y  H, We have (𝑥, (T1 T2 )∗ 𝑦) = ((T1 T2 )𝑥, 𝑦), {by definition of adjoint operator (T1 T2 ) }
= (T1 (T2 𝑥), 𝑦), {by composition of operators} Also we have T ∗ T = (T1 − 𝑖 T2 )(T1 + 𝑖 T2 ) = (T1 2 + T2 2 ) + 𝑖(T1 T2 − T2 T1 ), … (ii)
= (T2 𝑥, T1 ∗ (𝑦)), {by definition of adjoint operators} Now since T is a normal operator on H  TT ∗ = T ∗ T  (T1 2 + T2 2 ) + 𝑖( T2 T1 − T1 T2 ) = (T1 2 + T2 2 ) + 𝑖(T1 T2 − T2 T1 ), {using (i) and (ii)}
= (𝑥, T2 ∗ (T1 ∗ 𝑦)), {by definition of adjoint operators T1 & T2 }  ( T2 T1 − T1 T2 ) = (T1 T2 − T2 T1 ), {on equating real & imaginary parts}
∗
= (𝑥, (T2 ∗ T1 )𝑦), {by composition of operators }  2 T2 T1 = 2T1 T2 ,  T2 T1 = T1 T2 , This proves 1st part of the theorem
∗ ∗ Converse Part :- Assuming that the real part T1 and the imaginary part T2 of T commutes; i.e. T1 T2 = T2 T1 then we shall show that T is a normal
Thus we get (𝑥, (T1 T2 )∗ 𝑦) = (𝑥, (T2 ∗ T1 )𝑦)  (T1 T2 )∗ = T2 ∗ T1 , {by uniqueness of adjoint operator}. This proves (c)
operator:- Since T1 T2 = T2 T1 , then from (i) & (ii) we see that TT ∗ = T ∗ T  T is a normal operator on H. This completes the proof of the theorem.
Proof of (d):- Let any x, y  H, We have (x, T ∗∗ y) = (x, (T ∗ )∗ y), {by definition of adjoint operator T ∗ }
________________________________________________________________________________________________________________________
= (T ∗ 𝑥, y), {by definition of adjoint operator T ∗ }
___________________________________________________________________
̅̅̅̅̅̅̅̅̅̅
= (y, T ∗ 𝑥), {by property of inner product }
= ̅̅̅̅̅̅̅̅̅
(Ty, x), {by definition of adjoint operator }