The case of inner product spaces, in which we may provide a simpler condition of equality, is of interest in applications [6].
Theorem 6.4 (Dragomir, 2004). Let (X,k·k)be an inner product space over the real or complex number field K, ek, xi ∈ H\ {0}, k ∈ {1, . . . , m}, i ∈ {1, . . . , n}.IfMik ≥0fori∈ {1, . . . , n},{1, . . . , n}such that
(6.18) kxik −Rehxi, eki ≤Mik
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page40of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
for each i∈ {1, . . . , n}, k ∈ {1, . . . , m},then we have the inequality
(6.19)
n
X
i=1
kxik ≤
1 m
m
X
k=1
ek
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik.
The case of equality holds in (6.19) if and only if (6.20)
n
X
i=1
kxik ≥ 1 m
m
X
k=1 n
X
i=1
Mik
and (6.21)
n
X
i=1
xi = m Pn
i=1kxik −m1 Pm k=1
Pn
i=1Mik kPm
k=1ekk2
m
X
k=1
ek.
Proof. As in the proof of Theorem6.3, we have
(6.22)
n
X
i=1
kxik ≤Re
* 1 m
m
X
k=1
ek,
n
X
i=1
xi +
+ 1 m
m
X
k=1 n
X
i=1
Mik,
andPm
k=1ek6= 0.
On utilising the Schwarz inequality in the inner product space(H;h·,·i)for
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page41of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Pn
i=1xi,Pm
k=1ek,we have
n
X
i=1
xi
m
X
k=1
ek
≥
* n X
i=1
xi,
m
X
k=1
ek +
(6.23)
≥
Re
* n X
i=1
xi,
m
X
k=1
ek +
≥Re
* n X
i=1
xi,
m
X
k=1
ek +
.
By (6.22) and (6.23) we deduce (6.19).
Taking the norm in (6.21) and using (6.20), we have
n
X
i=1
xi
= m Pn
i=1kxik −m1 Pm k=1
Pn
i=1Mik kPm
k=1ekk ,
showing that the equality holds in (6.19).
Conversely, if the case of equality holds in (6.19), then it must hold in all the inequalities used to prove (6.19). Therefore we have
(6.24) kxik= Rehxi, eki+Mik for each i∈ {1, . . . , n}, k ∈ {1, . . . , m},
(6.25)
n
X
i=1
xi
m
X
k=1
ek
=
* n X
i=1
xi,
m
X
k=1
ek +
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page42of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
and
(6.26) Im
* n X
i=1
xi,
m
X
k=1
ek +
= 0.
From (6.24), on summing overiandk,we get
(6.27) Re
* n X
i=1
xi,
m
X
k=1
ek +
=m
n
X
i=1
kxik −
m
X
k=1 n
X
i=1
Mik.
On the other hand, by the use of the identity (3.22), the relation (6.25) holds if and only if
n
X
i=1
xi = hPn
i=1xi,Pm k=1eki kPm
k=1ekk2
m
X
k=1
ek,
giving, from (6.26) and (6.27), that
n
X
i=1
xi = mPn
i=1kxik −Pm k=1
Pn i=1Mik kPm
k=1ekk2
m
X
k=1
ek.
If the inequality holds in (6.19), then obviously (6.20) is valid, and the theorem is proved.
Remark 5. If in the above theorem the vectors {ek}k=1,m are assumed to be orthogonal, then (6.19) becomes:
(6.28)
n
X
i=1
kxik ≤ 1 m
m
X
k=1
kekk2
!12
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page43of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Moreover, if{ek}k=1,mis an orthonormal family, then (6.28) becomes
(6.29)
n
X
i=1
kxik ≤
√m m
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik,
which has been obtained in [12].
Before we provide some natural consequences of Theorem 6.4, we need some preliminary results concerning another reverse of Schwarz’s inequality in inner product spaces (see for instance [4, p. 27]).
Lemma 6.5 (Dragomir, 2004). Let(X,k·k)be an inner product space over the real or complex number field Kandx, a ∈ H, r > 0.Ifkx−ak ≤r, then we have the inequality
(6.30) kxk kak −Rehx, ai ≤ 1
2r2. The case of equality holds in (6.30) if and only if
(6.31) kx−ak=r and kxk=kak.
Proof. The conditionkx−ak ≤ris clearly equivalent to (6.32) kxk2+kak2 ≤2 Rehx, ai+r2. Since
(6.33) 2kxk kak ≤ kxk2+kak2,
with equality if and only if kxk =kak,hence by (6.32) and (6.33) we deduce (6.30).
The case of equality is obvious.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page44of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Utilising the above lemma we may state the following corollary of Theorem 6.4[6].
Corollary 6.6. Let (H;h·,·i), ek, xi be as in Theorem 6.4. If rik > 0, i ∈ {1, . . . , n}, k ∈ {1, . . . , m}such that
(6.34) kxi−ekk ≤rik for each i∈ {1, . . . , n} andk ∈ {1, . . . , m}, then we have the inequality
(6.35)
n
X
i=1
kxik ≤
1 m
m
X
k=1
ek
n
X
i=1
xi
+ 1 2m
m
X
k=1 n
X
i=1
rik2.
The equality holds in (6.35) if and only if
n
X
i=1
kxik ≥ 1 2m
m
X
k=1 n
X
i=1
r2ik
and n
X
i=1
xi = m Pn
i=1kxik − 2m1 Pm k=1
Pn i=1rik2 kPm
k=1ekk2
m
X
k=1
ek.
The following lemma may provide another sufficient condition for (6.18) to hold (see also [4, p. 28]).
Lemma 6.7 (Dragomir, 2004). Let (H;h·,·i)be an inner product space over the real or complex number fieldKandx, y ∈H, M ≥m >0.If either
(6.36) RehM y−x, x−myi ≥0
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page45of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
or, equivalently, (6.37)
x− m+M 2 y
≤ 1
2(M −m)kyk, holds, then
(6.38) kxk kyk −Rehx, yi ≤ 1
4· (M −m)2 m+M kyk2.
The case of equality holds in (6.38) if and only if the equality case is realised in (6.36) and
kxk= M +m 2 kyk.
The proof is obvious by Lemma6.5fora= M+m2 yandr= 12 (M −m)kyk. Finally, the following corollary of Theorem6.4may be stated [6].
Corollary 6.8. Assume that (H,h·,·i), ek, xi are as in Theorem6.4. IfMik ≥ mik >0satisfy the condition
RehMkek−xi, xi−µkeki ≥0 for each i∈ {1, . . . , n}andk ∈ {1, . . . , m},then
n
X
i=1
kxik ≤
1 m
m
X
k=1
ek
n
X
i=1
xi
+ 1 4m
m
X
k=1 n
X
i=1
(Mik−mik)2 Mik+mik
kekk2.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page46of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
7. Other Additive Reverses for m Functionals
A different approach in obtaining other additive reverses for the generalised triangle inequality is incorporated in the following new result:
Theorem 7.1. Let(X,k·k)be a normed linear space over the real or complex number field K. AssumeFk, k ∈ {1, . . . , m},are bounded linear functionals on the normed linear space X and xi ∈ X, i ∈ {1, . . . , n}, Mik ≥ 0, i ∈ {1, . . . , n}, k ∈ {1, . . . , m}are such that
(7.1) kxik −ReFk(xi)≤Mik for each i∈ {1, . . . , n} and k∈ {1, . . . , m}.
(i) Ifc∞is defined by (c∞), then we have the inequality
(7.2)
n
X
i=1
kxik ≤c∞
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik.
(ii) Ifcpis defined by (cp) forp≥1,then we have the inequality:
(7.3)
n
X
i=1
kxik ≤ 1 m1p
cp
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik.
Proof. (i) Since
1≤k≤mmax kFk(x)k ≤c∞kxk for anyx∈X,
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page47of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
then we have
m
X
k=1
Fk
n
X
i=1
xi
!
≤m max
1≤k≤m
Fk
n
X
i=1
xi
!
≤mc∞
n
X
i=1
xi .
Using (6.16), we may state that 1
m
m
X
k=1
ReFk n
X
i=1
xi
!
≤c∞
n
X
i=1
xi
,
which, together with (6.15) imply the desired inequality (7.2).
(ii) Using the fact that, obviously
m
X
k=1
|Fk(x)|p
!1p
≤cpkxk for anyx∈X, then, by Hölder’s inequality forp >1,1p +1q = 1,we have
m
X
k=1
Fk n
X
i=1
xi
!
≤m1q
m
X
k=1
Fk n
X
i=1
xi
!
p!1p
≤cpm1q
n
X
i=1
xi ,
which, combined with (6.15) and (6.16) will give the desired inequality (7.3).
The casep= 1goes likewise and we omit the details.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page48of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Remark 6. Since, obviouslyc∞≤ max
1≤k≤mkFkk,then from (7.2) we have
(7.4)
n
X
i=1
kxik ≤ max
1≤k≤m{kFkk} ·
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik.
Finally, sincecp ≤(Pm
k=1kFkkp)1p, p≥1,hence by (7.3) we have
(7.5)
n
X
i=1
kxik ≤ Pm
k=1kFkkp m
p1
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik.
The following corollary for semi-inner products may be stated as well.
Corollary 7.2. Let(X,k·k)be a real or complex normed space and[·,·] :X× X →Ka semi-inner product generating the normk·k.Assumeek, xi ∈H and Mik ≥0, i∈ {1, . . . , n}, k ∈ {1, . . . , m}are such that
(7.6) kxik −Re [xi, ek]≤Mik, for anyi∈ {1, . . . , n}, k ∈ {1, . . . , m}.
(i) If
d∞ := sup
x6=0
max1≤k≤n|[x, ek]|
kxk ≤ max
1≤k≤nkekk
,
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page49of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
then
n
X
i=1
kxik ≤d∞
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik (7.7)
≤ max
1≤k≤nkekk ·
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik
!
;
(ii) If
dp := sup
x6=0
Pm
k=1|[x, ek]|p kxkp
1p
≤
m
X
k=1
kekkp
!1p
,
wherep≥1,then
n
X
i=1
kxik ≤ 1 mp1
dp
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik (7.8)
≤ Pm
k=1kekkp m
1p
n
X
i=1
xi
+ 1 m
m
X
k=1 n
X
i=1
Mik
! .
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page50of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
8. Applications for Complex Numbers
LetCbe the field of complex numbers. Ifz = Rez+iImz,then by|·|p :C→ [0,∞), p∈[1,∞]we define thep−modulus ofz as
|z|p :=
max{|Rez|,|Imz|} if p=∞, (|Rez|p+|Imz|p)1p if p∈[1,∞), where|a|, a∈Ris the usual modulus of the real numbera.
Forp= 2,we recapture the usual modulus of a complex number, i.e.,
|z|2 = q
|Rez|2+|Imz|2 =|z|, z ∈C.
It is well known that
C,|·|p
, p ∈ [1,∞] is a Banach space over the real number field R.
Consider the Banach space(C,|·|1)andF :C→C,F (z) =azwitha∈C, a 6= 0.Obviously,F is linear onC. Forz 6= 0,we have
|F (z)|
|z|1 = |a| |z|
|z|1 = |a|
q
|Rez|2+|Imz|2
|Rez|+|Imz| ≤ |a|. Since, forz0 = 1,we have|F (z0)|=|a|and|z0|1 = 1,hence
kFk1 := sup
z6=0
|F (z)|
|z|1 =|a|,
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page51of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
showing thatF is a bounded linear functional on(C,|·|1)andkFk1 =|a|. We can apply Theorem3.1 to state the following reverse of the generalised triangle inequality for complex numbers [5].
Proposition 8.1. Letak, xj ∈ C,k ∈ {1, . . . , m}andj ∈ {1, . . . , n}.If there exist the constantsrk ≥0, k ∈ {1, . . . , m}withPm
k=1rk >0and (8.1) rk[|Rexj|+|Imxj|]≤Reak·Rexj −Imak·Imxj for eachj ∈ {1, . . . , n}andk ∈ {1, . . . , m},then
(8.2)
n
X
j=1
[|Rexj|+|Imxj|]≤ |Pm k=1ak| Pm
k=1rk
"
n
X
j=1
Rexj
+
n
X
j=1
Imxj
# .
The case of equality holds in (8.2) if both Re
m
X
k=1
ak
! Re
n
X
j=1
xj
!
−Im
m
X
k=1
ak
! Im
n
X
j=1
xj
!
=
m
X
k=1
rk
! n X
j=1
[|Rexj|+|Imxj|]
=
m
X
k=1
ak
"
n
X
j=1
Rexj
+
n
X
j=1
Imxj
# .
The proof follows by Theorem3.1applied for the Banach space(C,|·|1)and Fk(z) = akz, k ∈ {1, . . . , m}on taking into account that:
m
X
k=1
Fk 1
=
m
X
k=1
ak .
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page52of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Now, consider the Banach space (C,|·|∞).If F(z) = dz,then for z 6= 0 we have
|F (z)|
|z|∞ = |d| |z|
|z|∞ = |d|
q
|Rez|2 +|Imz|2 max{|Rez|,|Imz|} ≤√
2|d|.
Since, forz0 = 1 +i,we have|F (z0)|=√
2|d|,|z0|∞= 1,hence kFk∞:= sup
z6=0
|F (z)|
|z|∞ =√ 2|d|,
showing thatF is a bounded linear functional on(C,|·|∞)andkFk∞=√ 2|d|. If we apply Theorem 3.1, then we can state the following reverse of the generalised triangle inequality for complex numbers [5].
Proposition 8.2. Letak, xj ∈ C,k ∈ {1, . . . , m}andj ∈ {1, . . . , n}.If there exist the constantsrk ≥0, k ∈ {1, . . . , m}withPm
k=1rk >0and rkmax{|Rexj|,|Imxj|} ≤Reak·Rexj −Imak·Imxj for eachj ∈ {1, . . . , n}andk ∈ {1, . . . , m},then
(8.3)
n
X
j=1
max{|Rexj|,|Imxj|}
≤√
2·|Pm k=1ak| Pm
k=1rk max (
n
X
j=1
Rexj ,
n
X
j=1
Imxj
) .
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page53of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
The case of equality holds in (8.3) if both Re
m
X
k=1
ak
! Re
n
X
j=1
xj
!
−Im
m
X
k=1
ak
! Im
n
X
j=1
xj
!
=
m
X
k=1
rk
! n X
j=1
max{|Rexj|,|Imxj|}
=√ 2
m
X
k=1
ak
max (
n
X
j=1
Rexj ,
n
X
j=1
Imxj
) .
Finally, consider the Banach space
C,|·|2p
withp≥1.
LetF :C→C,F(z) = cz.By Hölder’s inequality, we have
|F (z)|
|z|2p = |c|
q
|Rez|2+|Imz|2
|Rez|2p+|Imz|2p2p1
≤212−2p1 |c|.
Since, forz0 = 1 +iwe have|F (z0)|= 212 |c|,|z0|2p = 22p1 (p≥1),hence kFk2p := sup
z6=0
|F (z)|
|z|2p = 212−2p1 |c|,
showing thatF is a bounded linear functional on
C,|·|2p
, p≥1andkFk2p = 212−2p1 |c|.
If we apply Theorem3.1, then we can state the following proposition [5].
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page54of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Proposition 8.3. Letak, xj ∈ C,k ∈ {1, . . . , m}andj ∈ {1, . . . , n}.If there exist the constantsrk ≥0, k ∈ {1, . . . , m}withPm
k=1rk >0and rk
|Rexj|2p+|Imxj|2p2p1
≤Reak·Rexj −Imak·Imxj
for eachj ∈ {1, . . . , n}andk ∈ {1, . . . , m},then
(8.4)
n
X
j=1
|Rexj|2p+|Imxj|2p2p1
≤212−2p1 |Pm k=1ak| Pm
k=1rk
n
X
j=1
Rexj
2p
+
n
X
j=1
Imxj
2p
1 2p
.
The case of equality holds in (8.4) if both:
Re
m
X
k=1
ak
! Re
n
X
j=1
xj
!
−Im
m
X
k=1
ak
! Im
n
X
j=1
xj
!
=
m
X
k=1
rk
! n X
j=1
|Rexj|2p+|Imxj|2p2p1
= 212−2p1
m
X
k=1
ak
n
X
j=1
Rexj
2p
+
n
X
j=1
Imxj
2p
1 2p
.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page55of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Remark 7. If in the above proposition we choose p = 1, then we have the following reverse of the generalised triangle inequality for complex numbers
n
X
j=1
|xj| ≤ |Pm k=1ak| Pm
k=1rk
n
X
j=1
xj
providedxj, ak, j ∈ {1, . . . , n},k ∈ {1, . . . , m}satisfy the assumption rk|xj| ≤Reak·Rexj−Imak·Imxj
for each j ∈ {1, . . . , n}, k ∈ {1, . . . , m}.Here |·| is the usual modulus of a complex number andrk>0, k ∈ {1, . . . , m}are given.
We can apply Theorem6.3 to state the following reverse of the generalised triangle inequality for complex numbers [6].
Proposition 8.4. Letak, xj ∈ C,k ∈ {1, . . . , m}andj ∈ {1, . . . , n}.If there exist the constantsMjk ≥0, k ∈ {1, . . . , m}, j∈ {1, . . . , n}such that
(8.5) |Rexj|+|Imxj| ≤Reak·Rexj −Imak·Imxj +Mjk for eachj ∈ {1, . . . , n}andk ∈ {1, . . . , m},then
(8.6)
n
X
j=1
[|Rexj|+|Imxj|]
≤ 1 m
m
X
k=1
ak
"
n
X
j=1
Rexj
+
n
X
j=1
Imxj
# + 1
m
m
X
k=1 n
X
j=1
Mjk.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page56of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
The proof follows by Theorem6.3applied for the Banach space(C,|·|1)and Fk(z) = akz, k ∈ {1, . . . , m}on taking into account that:
m
X
k=1
Fk 1
=
m
X
k=1
ak .
If we apply Theorem6.3 for the Banach space(C,|·|∞), then we can state the following reverse of the generalised triangle inequality for complex numbers [6].
Proposition 8.5. Letak, xj ∈ C,k ∈ {1, . . . , m}andj ∈ {1, . . . , n}.If there exist the constantsMjk ≥0, k ∈ {1, . . . , m}, j∈ {1, . . . , n}such that
max{|Rexj|,|Imxj|} ≤Reak·Rexj−Imak·Imxj+Mjk for eachj ∈ {1, . . . , n}andk ∈ {1, . . . , m},then
(8.7)
n
X
j=1
max{|Rexj|,|Imxj|}
≤
√2 m
m
X
k=1
ak
max (
n
X
j=1
Rexj ,
n
X
j=1
Imxj
) + 1
m
m
X
k=1 n
X
j=1
Mjk.
Finally, if we apply Theorem6.3, for the Banach space
C,|·|2p
withp≥1, then we can state the following proposition [6].
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page57of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Proposition 8.6. Letak, xj, Mjk be as in Proposition8.5. If
|Rexj|2p+|Imxj|2p2p1
≤Reak·Rexj−Imak·Imxj +Mjk for eachj ∈ {1, . . . , n}andk ∈ {1, . . . , m},then
(8.8)
n
X
j=1
|Rexj|2p+|Imxj|2p2p1
≤ 212−2p1 m
m
X
k=1
ak
n
X
j=1
Rexj
2p
+
n
X
j=1
Imxj
2p
1 2p
+ 1 m
m
X
k=1 n
X
j=1
Mjk.
wherep≥1.
Remark 8. If in the above proposition we choose p = 1, then we have the following reverse of the generalised triangle inequality for complex numbers
n
X
j=1
|xj| ≤
1 m
m
X
k=1
ak
n
X
j=1
xj
+ 1 m
m
X
k=1 n
X
j=1
Mjk
providedxj, ak, j ∈ {1, . . . , n},k ∈ {1, . . . , m}satisfy the assumption
|xj| ≤Reak·Rexj −Imak·Imxj+Mjk
for each j ∈ {1, . . . , n}, k ∈ {1, . . . , m}.Here |·| is the usual modulus of a complex number andMjk >0, j ∈ {1, . . . , n},k ∈ {1, . . . , m}are given.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page58of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
9. Karamata Type Inequalities in Hilbert Spaces
Let f : [a, b] → K, K = C or R be a Lebesgue integrable function. The following inequality, which is the continuous version of the triangle inequality (9.1)
Z b a
f(x)dx
≤ Z b
a
|f(x)|dx,
plays a fundamental role in Mathematical Analysis and its applications.
It appears, see [20, p. 492], that the first reverse inequality for (9.1) was obtained by J. Karamata in his book from 1949, [14]. It can be stated as
(9.2) cosθ
Z b a
|f(x)|dx≤
Z b a
f(x)dx provided
−θ ≤argf(x)≤θ, x∈[a, b]
for givenθ ∈ 0,π2 .
This result has recently been extended by the author for the case of Bochner integrable functions with values in a Hilbert spaceH (see also [10]):
Theorem 9.1 (Dragomir, 2004). If f ∈ L([a, b] ;H) (this means that f : [a, b]→His strongly measurable on[a, b]and the Lebesgue integralRb
a kf(t)kdt is finite), then
(9.3)
Z b a
kf(t)kdt≤K
Z b a
f(t)dt ,
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page59of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
provided thatf satisfies the condition
(9.4) kf(t)k ≤KRehf(t), ei for a.e. t∈[a, b], wheree∈H,kek= 1andK ≥1are given.
The case of equality holds in (9.4) if and only if (9.5)
Z b a
f(t)dt = 1 K
Z b a
kf(t)kdt
e.
As some natural consequences of the above results, we have noticed in [10]
that, ifρ∈[0,1)andf ∈L([a, b] ;H)are such that (9.6) kf(t)−ek ≤ρ for a.e. t∈[a, b], then
(9.7) p
1−ρ2 Z b
a
kf(t)kdt ≤
Z b a
f(t)dt with equality if and only if
Z b a
f(t)dt=p 1−ρ2
Z b a
kf(t)kdt
·e.
Also, foreas above and ifM ≥m >0, f ∈L([a, b] ;H)such that either (9.8) RehM e−f(t), f(t)−mei ≥0
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page60of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
or, equivalently, (9.9)
f(t)− M +m 2 e
≤ 1
2(M −m) for a.e. t ∈[a, b],then
(9.10)
Z b a
kf(t)kdt ≤ M+m 2√
mM
Z b a
f(t)dt ,
with equality if and only if Z b
a
f(t)dt = 2√ mM M +m
Z b a
kf(t)kdt
·e.
The main aim of the following sections is to extend the integral inequalities mentioned above for the case of Banach spaces. Applications for Hilbert spaces and for complex-valued functions are given as well.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page61of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
10. Multiplicative Reverses of the Continuous Triangle Inequality
10.1. The Case of One Functional
Let (X,k·k) be a Banach space over the real or complex number field. Then one has the following reverse of the continuous triangle inequality [11].
Theorem 10.1 (Dragomir, 2004). Let F be a continuous linear functional of unit norm onX.Suppose that the functionf : [a, b]→Xis Bochner integrable on[a, b]and there exists ar≥0such that
(10.1) rkf(t)k ≤ReF [f(t)] for a.e. t∈[a, b]. Then
(10.2) r
Z b a
kf(t)kdt ≤
Z b a
f(t)dt ,
where equality holds in (10.2) if and only if both
(10.3) F
Z b a
f(t)dt
=r Z b
a
kf(t)kdt
and
(10.4) F
Z b a
f(t)dt
=
Z b a
f(t)dt .
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page62of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Proof. Since the norm ofF is one, then
|F(x)| ≤ kxk for any x∈X.
Applying this inequality for the vectorRb
af(t)dt,we get
Z b a
f(t)dt
≥ F
Z b a
f(t)dt
(10.5)
≥
ReF Z b
a
f(t)dt
=
Z b a
ReF (f(t))dt .
Now, by integration of (10.1), we obtain (10.6)
Z b a
ReF (f(t))dt ≥r Z b
a
kf(t)kdt,
and by (10.5) and (10.6) we deduce the desired inequality (10.2).
Obviously, if (10.3) and (10.4) hold true, then the equality case holds in (10.2).
Conversely, if the case of equality holds in (10.2), then it must hold in all the inequalities used before in proving this inequality. Therefore, we must have (10.7) rkf(t)k= ReF(f(t)) for a.e. t∈[a, b],
(10.8) ImF
Z b a
f(t)dt
= 0
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page63of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
and (10.9)
Z b a
f(t)dt
= ReF Z b
a
f(t)dt
.
Integrating (10.7) on[a, b],we get
(10.10) r
Z b a
kf(t)kdt= ReF Z b
a
f(t)dt
.
On utilising (10.10) and (10.8), we deduce (10.3) while (10.9) and (10.10) would imply (10.4), and the theorem is proved.
Corollary 10.2. Let (X,k·k)be a Banach space,[·,·] : X×X → Ra semi-inner product generating the norm k·kande ∈ X, kek = 1.Suppose that the functionf : [a, b]→X is Bochner integrable on[a, b]and there exists ar ≥ 0 such that
(10.11) rkf(t)k ≤Re [f(t), e] for a.e. t∈[a, b]. Then
(10.12) r
Z b a
kf(t)kdt ≤
Z b a
f(t)dt where equality holds in (10.12) if and only if both (10.13)
Z b a
f(t)dt, e
=r Z b
a
kf(t)kdt
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page64of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
and (10.14)
Z b a
f(t)dt, e
=
Z b a
f(t)dt .
The proof follows from Theorem 10.1 for the continuous linear functional F (x) = [x, e], x∈X,and we omit the details.
The following corollary of Theorem10.1may be stated [8].
Corollary 10.3. Let (X,k·k) be a strictly convex Banach space, [·,·] : X × X → Ka semi-inner product generating the normk·kande ∈X,kek = 1.If f : [a, b]→Xis Bochner integrable on[a, b]and there exists ar≥0such that (10.11) holds true, then (10.12) is valid. The case of equality holds in (10.12) if and only if
(10.15)
Z b a
f(t)dt=r Z b
a
kf(t)kdt
e.
Proof. If (10.15) holds true, then, obviously
Z b a
f(t)dt
=r Z b
a
kf(t)kdt
kek=r Z b
a
kf(t)kdt,
which is the equality case in (10.12).
Conversely, if the equality holds in (10.12), then, by Corollary10.2, we must have (10.13) and (10.14). Utilising Theorem4.2, by (10.14) we can conclude that there exists aµ >0such that
(10.16)
Z b a
f(t)dt =µe.
Reverses of the Triangle Inequality in Banach Spaces
S.S. Dragomir
Title Page Contents
JJ II
J I
Go Back Close
Quit Page65of99
J. Ineq. Pure and Appl. Math. 6(5) Art. 129, 2005
http://jipam.vu.edu.au
Replacing this in (10.13), we get µkek2 =r
Z b a
kf(t)kdt,
giving
(10.17) µ=r
Z b a
kf(t)kdt.
Utilising (10.16) and (10.17) we deduce (10.15) and the proof is completed.