1/33
Looks like no tags are added yet.
Subspace: what 3 things do you check?
(1) 0 ∈ U. (2) Closed under addition: u, v ∈ U ⟹ u + v ∈ U. (3) Closed under scalar multiplication: λ ∈ F, u ∈ U ⟹ λu ∈ U.
Quickest way to show a set is NOT a subspace?
Show the zero vector isn't in it, or that it fails closure. (The zero VECTOR is not the empty set.)
Linear combination of v1, …, vk ::
: a1v1 + … + akvk with a1, …, ak ∈ F.
Span(v1, …, vk) :::
The set of all linear combinations. It is a subspace, and the smallest subspace containing v1, …, vk.
What does 'v1, …, vk spans V' mean? :::
span(v1, …, vk) = V.
Finite-dimensional :::
V has some finite list that spans it.
Linearly dependent :::
Not independent: there is a solution to a1v1 + … + akvk = 0 with some ai ≠ 0.
Basis ::
: A list that is linearly independent AND spans V.
Dimension :::
The length of any basis of V (well-defined because all bases have the same length).
U + W :::
{u + w : u ∈ U, w ∈ W}. The smallest subspace containing both U and W.
Direct sum U ⊕ W (definition) :::
U + W where every element can be written as u + w in only one way.
V = U ⊕ W means which two things? :::
U + W = V, and U ∩ W = {0}.
Standard basis of Fⁿ and of Pn(F) :
:: Fⁿ: (1,0,…,0), …, (0,…,0,1). Pn(F): 1, x, x², …, xⁿ.
What is Pn(F)? :
:: Polynomials of degree at most n with coefficients in F. dim Pn(F) = n + 1.
Linear dependence lemma (2.19) :::
If v1, …, vm is dependent, then some vj is in span(v1, …, v(j-1)), and removing it doesn't change the span.
Length of independent list vs spanning list (2.22)
::: Length of any independent list ≤ length of any spanning list.
Why do all bases have the same length? (2.34) ::
: Take two bases B1, B2. B1 independent and B2 spans, so |B1| ≤ |B2|. Swap roles, so |B2| ≤ |B1|. Hence equal.
Unique representation in a basis :::
If v1, …, vn is a basis, every v ∈ V is a1v1 + … + anvn for exactly one choice of a's.
If dim V = n, what shortcut do you have for bases? :::
Any independent list of length n is a basis, and any spanning list of length n is a basis.
If U ⊆ V (V finite-dimensional), compare dimensions
::: dim U ≤ dim V, and dim U = dim V ⟹ U = V.
Extending to a basis (2.32) :::
Every independent list in a finite-dimensional V can be extended to a basis of V.
Complement of a subspace :
:: If U ⊆ V, there is W with V = U ⊕ W. (Extend a basis of U to a basis of V and let W be the span of the added vectors.)
Dimension of a sum (2.43) ::
: dim(U + W) = dim U + dim W − dim(U ∩ W).
U + W is direct if and only if…
::: U ∩ W = {0}.
Recipe: basis of a subspace defined by equations
::: (1) Solve the equations in terms of free variables. (2) Write the general vector as free1·vector + free2·vector + … (3) Those vectors span U. (4) Show they're independent. (5) Conclude it's a basis; dim U = number of vectors.
Recipe: show a list is linearly independent :
:: Write a1v1 + … + akvk = 0, compare coordinates (or coefficients) to get equations, solve, and show all a's must be 0.
Recipe: extend a basis of U to a basis of V :
:: Add vectors until the list has length dim V. Check independence. Independent + length = dim V ⟹ basis.
Recipe: find W with V = U ⊕ W :::
W = span of the vectors added when extending. Justify: every vector is a unique combination of the extended basis, so it splits uniquely as u + w.
In a justification, what must you always include for 'is this a basis'? ::
: Both spanning AND linear independence (or independence plus length = dim V). Missing one loses marks.
Prove U ∩ W is a subspace :
:: (1) 0 ∈ U and 0 ∈ W, so 0 ∈ U ∩ W. (2) x, y ∈ U ∩ W ⟹ x + y ∈ U and x + y ∈ W ⟹ x + y ∈ U ∩ W. (3) x ∈ U ∩ W, λ ∈ F ⟹ λx ∈ U and λx ∈ W ⟹ λx ∈ U ∩ W.
Prove: U + W direct ⟹ U ∩ W = {0} (from the uniqueness definition) :::
Take v ∈ U ∩ W. Then v = v + 0 (v ∈ U, 0 ∈ W) and v = 0 + v (0 ∈ U, v ∈ W). Two ways to write v, so by uniqueness v = 0.
Prove: U ∩ W = {0} ⟹ U + W direct ::
: Suppose u + w = u' + w'. Then u − u' = w' − w. Left side ∈ U, right side ∈ W, so both ∈ U ∩ W = {0}. So u = u' and w = w'.
Idea of the proof of 2.22 :::
Start with a spanning list w1, …, wn. Add u1 at the front: now dependent, so by the lemma some w can be removed and the list still spans. Repeat with u2, u3, …. Each time a u goes in, a w goes out, and the removed vector can never be a u (they're independent). So you never run out of w's, hence m ≤ n.
Proof of 2.34 :::
B1 independent, B2 spans ⟹ |B1| ≤ |B2| (by 2.22). Swap roles ⟹ |B2| ≤ |B1|. So they're equal.