Suppose V is finite-dimensional and T1,T2∈L(V,W). Prove that range T1=range T2 if and only if there exists an invertible operator S∈L(V) such that T1=T2S.
(In other words, find an isomorphism S that reshuffles the inputs so T1=T2S)
Suppose range T1=range T2.
Notice the nullspaces have the same dimension by 3.22 and the ranges being equal, meaning we can let u1,…,un be a basis for null T1 and u1′,…,un′ be a basis of null T2 (notice they're both of length n).
Extend u1,…,un to a basis u1,…,un,v1,…,vr of V.
Because the ranges are equal we can define v1′,…,vr′ such that T1vk=T2vk′.
To see that u1′,…,un′,v1′,…,vr′ is a basis of V we need only show it's independent, since it's the right length 2.39 will imply it's a basis. (Note this argument is similar to the proof of 3.22)
⟹⟹⟹a1v1′+⋯+arvr′+c1u1′+⋯+cnun′=0T1(a1v1+⋯+arvr)=0a1v1+⋯+arvr=b1u1+⋯+bnunak=bk=0apply T2 and use T2vk′=T1vkas a1v1+⋯+arvr∈ null T1as u1,…,un,v1,…,vr is indep
Thus we've shown ak=0, meaning the original equation becomes c1u1′+⋯+cnun′=0
Which obviously implies ck=0 since u1′,…,un′ is a basis for null T2.
Define Suk=uk′ and Svk=vk′. S is clearly an isomorphism (with inverse S−1uk′=uk, S−1vk′=vk) as we're just moving from one basis of V to another.
We have T1=T2S because they agree on the basis vectors (T1uk=T2Suk and T1vk=T2Svk). This completes the forward direction.
Now suppose T1=T2S. If w=T1v then w=T2(Sv) implying range T1⊆range T2 likewise if w=T2v then w=T1(S−1v) implying range T2⊆range T1.
todo: diagram / graphic for the proof idea (what linear map looks like). also describe the general strategy of preranges, postnullspaces etc and how you construct them