Click here to load reader
Upload
vinit-garg
View
214
Download
0
Embed Size (px)
Citation preview
8/9/2019 field1
http://slidepdf.com/reader/full/field1 1/4
AM 106/206: Applied Algebra Prof. Salil Vadhan
Lecture Notes 20
November 22, 2010
1 Vector Spaces
• Reading: Gallian Ch. 19
• Today’s main message: linear algebra (as in Math 21) can be done over any field, and mostof the results you’re familiar with from the case of R or C carry over.
• Def: A vector space over a field F is a set V with two operations + : V × V → V (vectoraddition) and · : F × V → V (scalar multiplications) that satisfy the following properties:
1. V is an abelian group under +.2. (ab) · v = a · (b · v) for all a, b ∈ F and v ∈ V .
3. 1 · v = v for all v ∈ V .
4. a · (v + w) = a · v + a · w for all a ∈ F and v , w ∈ V .
5. (a + b) · v = a · v + b · v for all a, b ∈ F and v ∈ V .
• A vector space has more structure than an abelian group, but less structure than a ring (onlymultiplication by scalars, not multiplication of arbitrary pairs of elements of V ).
• Examples and Nonexamples:
– V = F n
– V = C, F = R
– V = Zn, F = Z2
– V = F [x]
– V = F [x]/ p(x)
– V = R for a ring R containing F .
1
8/9/2019 field1
http://slidepdf.com/reader/full/field1 2/4
8/9/2019 field1
http://slidepdf.com/reader/full/field1 3/4
• Prop: Every n-dimensional vector space V over F is isomorphic to F n.
Proof: Let v1, . . . , vn be a basis for V .Then an isomorphism from F n to V is given by:
• Matrices: A linear map f : F n → F m can be described uniquely by an m × n matrix M with entries from F .
– M ij = f (e j)i, where e j = (000 · · · 010 · · · 00) has a 1 in the j ’th position.
– For v = (v1, . . . , vn) ∈ F n, f (v)i = f (
j v je j)i =
j v jf (e j)i =
i M ijv j = (M v)i,where M v is matrix-vector product.
– Matrix multiplication ↔ composition of linear maps.
– If n = m, then f is an isomorphism ⇔ det(M ) = 0.
– Solving M v = w for v (when given M and w ∈ F m) is equivalent to solving a linearsystem with m variables and n unknowns.
• Example: f : Z33 → Z3
2 given by f (v1, v2, v3) = (v1 + 2v2, 2v1 + v3).
• Thm: If f : V → W is a linear map, then dim(ker(f )) + dim(im(f )) = dim(V ).
Proof: omitted.
• When F finite, this says |V | = |F |dim(V ) = |F |dim(ker(f )) · |F |dim(im(f )) = | ker(f )| · |im(f )|, justlike for group homomorphisms!
• Corollaries:
– F n F m if m = n.
– All bases of a vector space have the same size.
– A homogenous linear system M v = 0 for a given m × n matrix M always has a nonzerosolution v if n > m (more variables than unknowns).
• Computational issues: For n × n matrices over F ,
– Matrix multiplication can be done with O(n3) operations in F using the standard algo-rithm.
– The determinant and inverse, and solving a linear system M v = w can be done usingO(n3) operations in F using Gaussian elimination. (For infinite fields, need to worryabout the size of the numbers, or accuracy if doing approximate arithmetic. No suchproblem in finite fields.)
3
8/9/2019 field1
http://slidepdf.com/reader/full/field1 4/4
– Asymptotically fastest known algorithms run in time O(n2.376). Whether time O(n2) ispossible is a long-standing open problem.
3 Application to Extension Fields
• Reading: parts of Gallian Ch. 21
• Def: E is an extension field of F is F is a subfield of E . The degree of E over F is thedimension of E as a vector space over F , and is denoted [E : F ]. E is a finite extension if [E : F ] is finite.
• Examples:
– [C : R] =
– [F [x]/ p(x) : F ] =
– [F(α) : F] =
• Thm 21.5: If K is a finite extension of E , and E is a finite extension of F , then [K : F ] =[K : E ][E : F ].
• Example: The splitting field of x8 − 1 over Q, i.e. Q(ω) = Q(i)(ω), where ω = e2πi/8.
• Proof:
• Corollary: If E is a finite extension of F and α ∈ E , then α is algebraic over F and itsminimal polynomial has degree dividing [E : F ].
4