Scribd
Upload
6 views7 pages

Shiblu Linear

Uploaded by

Shahariar Nahid
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views7 pages

Shiblu Linear

Uploaded by

Shahariar Nahid
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

AssignmentTopic: SUMS AND DIRECT SUMS

Course Code:MAT-2211
Course Name: Linear Algebra
Submitted To Submitted By
Mamun-Or-Rashid Shiblu Nath
Lecturer, Semester:3rd
Dept.Of Mathematics ID:0222220012201008
Premier University

Submission Date :7/11/24


SUMS AND DIRECT SUMS
Definition (SUMS)
Suppose U1,...,Um are subsets of V. The sum of U1,...,Um denoted
by U1+...+Um is the set of all possible sums of elements U1,...,Um.
More precisely,
U1+...+Um={u1+...+um:u1∈U1,...,um∈Um}
Definition (DIRECT SUMS)
Suppose U1,...,Um are subspaces of V.
The sum U1+...+Um is called a direct sum if and only if each
element of U1+...+Um can be written in one and only one way as a
sum of u1+...+um where each uj is in Uj. The notation for such a
sum of subsets is U1⊕...⊕Um.
Example:
Consider the vector space V=𝑅 3
(a)= Let U be the xy-plane and let W be the yz-plane ;that is
U={(a,b,0):a,b ∈ 𝑅} and W=={(0,b,c):b,c ∈ 𝑅}
Then 𝑅 3 =U+W ,since every vector in
𝑅 3 𝑖𝑠 𝑡ℎ𝑒 𝑠𝑢𝑚 𝑜𝑓 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟 𝑖𝑛 𝑈 𝑎𝑛𝑑 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟 𝑖𝑛 𝑊. 𝐻𝑜𝑤𝑒𝑣𝑒𝑟, 𝑅 3 𝑖𝑠 𝑛𝑜𝑡
𝑡ℎ𝑒 𝑑𝑖𝑟𝑒𝑐𝑡 𝑠𝑢𝑚 𝑜𝑓 𝑈 and W,since such sums are not unique .For
example (3,5,7)=(3,1,0)+(0,4,7) and also (3,5,7)=(3,-4,0)+(0,9,7)
(b)=Let U be the xy-p;ane and let W be the z- axis ,That is
U={(a,b,0):a,b∈ 𝑅} and W={(0,0,c):c∈ 𝑅 }
Now any vector (a,b,c) ∈
𝑅 3 𝑐𝑎𝑛 𝑏𝑒 𝑤𝑟𝑖𝑡𝑡𝑒𝑛 𝑎𝑠 𝑡ℎ𝑒 𝑠𝑢𝑚 𝑜𝑓 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟 𝑖𝑛 𝑈 𝑎𝑛𝑑 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟 𝑖𝑛 𝑉 𝑖𝑛 𝑜𝑛𝑒 𝑎𝑛𝑑
𝑜𝑛𝑙𝑦 𝑤𝑎𝑦; (a,b,c)=(a,b,0)+(0,0,c)
Accordingly , 𝑅 3 𝑖𝑠 𝑡ℎ𝑒 𝑑𝑖𝑟𝑒𝑐𝑡 𝑠𝑢𝑚 𝑜𝑓 𝑈 𝑎𝑛𝑑 𝑊 ; 𝑡ℎ𝑎𝑡 𝑖𝑠 𝑅 3 =U⊕W

Theorem :

If 𝑤1 and 𝑤2 are subspaces of a vector space V ,then


dim(𝑤1 +𝑤2 )=dim𝑤1 +dim𝑤2 -dim(𝑤1 ∩𝑤2 )
Proof:

Let S be a basis of W1∩W2 ( if 𝑤1 ∩𝑤2 is the zero space then S= ∅).For


each i=1 , 2 extend S to a basis 𝐵𝑖 of 𝑤𝑖 .Let S={𝑢1 , 𝑢2,… 𝑢𝑟, },𝐵1, ={
𝑢1 , 𝑢2 …,𝑢𝑟, , 𝑣1, 𝑣2 , … , 𝑣𝑠 } and
𝐵2 ={𝑢1 , 𝑢2,… 𝑢𝑟, 𝑤1, 𝑤2, … , 𝑤𝑡, }Then dim(𝑤1 ∩𝑤2 )=r,
dim𝑤1 = 𝑟 + 𝑠, dim𝑤2 = 𝑟 + 𝑡, 𝐿𝑒𝑡 𝐵 =
{𝑢1 , 𝑢2,… 𝑢𝑟, 𝑣1, 𝑣2 , … , 𝑣𝑠 , 𝑤1, 𝑤2, … , 𝑤𝑡, }..It is enough to show
that B is a basis of 𝑤1 +𝑤2 because then we have
dim(𝑤1 +𝑤2 )=r+s+t= (r+s)+(r+t)-r= dim𝑤1 +dim𝑤2 -dim(𝑤1 ∩𝑤2 ).
To show that B is linearly independent.
𝑟 𝑠
Let , ∑𝑖=0 𝑎𝑖 𝑢𝑖 + ∑ 𝑏𝑗 𝑣𝑗 + ∑𝑡𝑘=1 𝑐𝑘 𝑤𝑘 =0
𝑗=1

Then ,
𝑠
𝑟 𝑡

∑ 𝑎𝑖 𝑢𝑖 + ∑ 𝑏𝑗 𝑣𝑗 = − ∑ 𝑐𝑘 𝑤𝑘
𝑖=0 𝑘=1
𝑗=1

Now the LHS is in 𝑤1 and the RHS is in


𝑤2 . 𝑆𝑜 𝑡ℎ𝑖𝑠 𝑒𝑙𝑒𝑚𝑒𝑛𝑡 𝑖𝑠 𝑖𝑛 𝑤1 ∩𝑤2 .Thus − ∑𝑡𝑘=1 𝑐𝑘 𝑤𝑘 =
∑𝑡𝑘=1 𝑑𝑖 𝑢𝑖 . 𝑠𝑜 𝑡ℎ𝑎𝑡 ∑𝑟𝑖=1 𝑑𝑖 𝑢𝑖 + ∑𝑡𝑘=1 𝑐𝑘 𝑤𝑘 =0 which implies
𝑑𝑖 =0and𝑐𝑘 =
0 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑖 𝑎𝑛𝑑 𝑘 (𝑠𝑖𝑛𝑐𝑒𝐵2 𝑖𝑠 𝑙𝑖𝑛𝑒𝑎𝑟𝑙𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 ).

𝑠
𝑇ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒 ∑𝑟𝑖=0 𝑎𝑖 𝑢𝑖 +∑ 𝑏𝑗 𝑣𝑗 = 0 which implies 𝑎𝑖 =0
𝑗=0
and 𝑏𝑗 =
0 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑖 𝑎𝑛𝑑 𝑗 (𝑠𝑖𝑛𝑐𝑒𝐵1 𝑖𝑠 𝑙𝑖𝑛𝑒𝑎𝑟𝑙𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 )

Thus B is linearly independent .Let w ∈ 𝑤1 +𝑤2 .Then w=𝑤1 +𝑤2 for


some
𝑤𝑖 ∈ 𝑊𝑖 for i=1,2. Then 𝑤𝑖 = ∑𝑟𝑖=0 𝑝𝑖 𝑢𝑖 +
𝑠
∑ 𝑞𝑗 𝑣𝑗 𝑎𝑛𝑑 𝑤2 = ∑𝑟𝑖=0 𝑔𝑖 𝑢𝑖 +
𝑗=0
𝑡
𝑟
∑ ℎ𝑗 𝑤𝑗 𝑓𝑜𝑟 𝑝𝑖 , 𝑞𝑖 , 𝑔𝑖 , ℎ𝑖 ∈F. Now w=∑ (𝑝
𝑖=1 𝑖
+ 𝑔𝑖 )
𝑗=1
𝑠
𝑡
𝑢𝑖 + ∑ 𝑞𝑗 𝑣𝑗 +∑𝑘=1 ℎ𝑘 𝑤𝑘 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑖𝑛 𝑠𝑝𝑎𝑛 (𝐵)
𝑗=1

Hence we can prove that dim(𝑤1 +𝑤2 )=dim𝑤1 +dim𝑤2 -


dim(𝑤1 ∩𝑤2 )

dim(𝑤1 +𝑤2 )=dim𝑤1 +dim𝑤2 -dim(𝑤1 ∩𝑤2 ) [Proved]

Theorem :
If V be a a finite dimensional vector space which direct sum of its two
subspace U and W then prove that dimV=dimU+dimW
Proof:
Suppose that dimV=n. Since it is finite dimensional then the
dimension of U and W be also finite .Let dimU=m≤n;and dimw=
l≤n. Let { 𝑢1 , 𝑢2 … ..,𝑢𝑚 } will be the basis of u and { 𝑤1 , 𝑤2 ,…… 𝑤𝑙 }
will be the basis of W.And suppose that ,B={ 𝑢1 , 𝑢2 … ..,𝑢𝑚 𝑤1 ,
𝑤2 ,…… 𝑤𝑙 } .Ww will show that B is a basis of V.
Since B contains exactly (m+n) number of elements and if B is a
basis of V then dimV= m+l Now v=u⊕v then for any
v∈V
i)V={ u+w u∈U and and v∈V}
ii) U∩W ={0}
Since u ∈U and {𝑢1 , 𝑢2 … ..,𝑢𝑚 } 𝑖𝑠 𝑎 𝑏𝑎𝑠𝑖𝑠 𝑜𝑓 𝑈 𝑡ℎ𝑒𝑛 𝑈 =
𝑎1 𝑢1+𝑎2 𝑢2 + ⋯ + 𝑎𝑚 𝑢𝑚 , also if w∈W and { 𝑤1 , 𝑤2 ,…… 𝑤𝑙 }
is a basis of then W = 𝑏1 𝑤1 +𝑏2 𝑤2 + ⋯ + 𝑏𝑙 𝑤𝑙
By conditions (i)
V=u+w
V=𝑎1 𝑢1 +𝑎2 𝑢2 + ⋯ + 𝑎𝑚 𝑢𝑚 + 𝑏1 𝑤1 +𝑏2 𝑤2 + ⋯ + 𝑏𝑙 𝑤 𝑙..

as v∈V V is linear combination of {𝑢1 , 𝑢2 … ..,𝑢𝑚 , 𝑤1 ,


𝑤2 ,…𝑤𝑙 }𝑖. 𝑒 𝑉 𝑖𝑠 𝑎 𝑙𝑖𝑛𝑒𝑎𝑟 𝑐𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑙𝑒𝑚𝑒𝑛𝑡 𝑜𝑓 𝐵 𝑖. 𝑒 𝑉 =
𝐿(𝐵) 𝑁𝑒𝑥𝑡 𝑤𝑒 𝑤𝑖𝑙𝑙 𝑠ℎ𝑜𝑤 𝑡ℎ𝑎𝑡 𝐵 𝑖𝑠 𝑙𝑖𝑛𝑒𝑎𝑟𝑙𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 .

𝐿𝑒𝑡, 𝑎1 𝑢1 +𝑎2 𝑢2 + ⋯ + 𝑎𝑚 𝑢𝑚 + 𝑏1 𝑤1 +𝑏2 𝑤2 + ⋯ +


𝑏𝑙 𝑤𝑙 = 0
𝑎1 𝑢1 +𝑎2 𝑢2 + ⋯ + 𝑎𝑚 𝑢𝑚 = −𝑏1 𝑤1 -𝑏2 𝑤2 − ⋯ − 𝑏𝑙 𝑤𝑙 =
0
Or, ∑𝑚 𝑙
𝑖=1 𝑎𝑖 𝑢𝑖 =− ∑𝑗=1 𝑏 𝑤𝑗𝑗

∑𝑚 𝑙
𝑖=1 𝑎𝑖 𝑢𝑖 ∈U and ∑𝑗=1 𝑏 𝑤𝑗 ∈W or, ∑𝑚 𝑙
𝑖=1 𝑎𝑖 𝑢𝑖 ∑𝑗=1 𝑏 𝑤𝑗 ∈U and
𝑗 𝑗
∑𝑚 𝑙 𝑚 𝑙
𝑖=1 𝑎𝑖 𝑢𝑖 ∑𝑗=1 𝑏 𝑤𝑗 ∈W or, ∑𝑖=1 𝑎𝑖 𝑢𝑖 ∑𝑗=1 𝑏 𝑤𝑗 ∈U∩W
𝑗 𝑗

But by condition (ii) U∩W={0} which shows that ∑𝑚


𝑖=1 𝑎𝑖 𝑢𝑖 =

0 𝑎𝑛𝑑 ∑𝑙𝑗=1 𝑏 𝑤𝑗 = 0 𝑖. 𝑒 𝑎1 𝑢1+𝑎2 𝑢2 + ⋯ + 𝑎𝑚 𝑢𝑚 =


𝑗
0 𝑎𝑛𝑑 𝑏1 𝑤1 +𝑏2 𝑤2 + ⋯ + 𝑏𝑙 𝑤𝑙 = 0 Since {𝑢𝑖 }<
{𝑤𝑗 } 𝑖𝑠 𝑎 𝑏𝑎𝑠𝑖𝑠 𝑜𝑓 𝑢 𝑎𝑛𝑑 𝑤 𝑟𝑒𝑠𝑝𝑒𝑐𝑡𝑖𝑣𝑒𝑙𝑦 𝑡ℎ𝑎𝑡 ′ 𝑠 𝑤ℎ𝑦 𝑡ℎ𝑒𝑦 𝑎𝑟𝑒 𝐿. 𝐼 𝑆𝑜
, 𝑎1 𝑢1 +𝑎2 𝑢2 + ⋯ + 𝑎𝑚 𝑢𝑚 = 0 𝑜𝑟 𝑎1 = 𝑎2 = ⋯ =
𝑎𝑚 = 0 𝑎𝑛𝑑 𝑏1 𝑤1 +𝑏2 𝑤2 + ⋯ + 𝑏𝑙 𝑤𝑙 = 0 or, 𝑏1 = 𝑏2 =
⋯ = 𝑏𝑙 =0 Since 𝑎1 = 𝑎2 = ⋯ = 𝑎𝑚 = 𝑏1 = 𝑏2 = ⋯ =
𝑏𝑙 = 0 So the vectors {𝑢𝑖 , 𝑤𝑗 } are L.I i.e B is linearly independent .Since
B is L.I and V =L(B).So B is the basis of V .
Here dimV=dimU+dimW [Proved]

You might also like