Midterm Review
Geometric Algorithms
Boston University
Fall 2023

This a collection of things you should be comfortable with going into the midterm. It is not exhaustive, and I can pretty much guarantee that no question on the midterm will be a direct restatement of anything below, but if you understand everything in this document, you have the foundations to approach any question on the midterm. You will have to add to it as you see fit.

Objects you should know

Things you should know how to do

Facts you should know how to use

  • every matrix has a unique reduced echelon form
  • a matrix is the augmented matrix of a inconsistent system if and only if any of its echelon forms have a pivot in the last column
  • two vectors are linearly dependent if and only if they are co-linear
  • the following are equivalent for a (m × n) matrix A:
    • Ax = 0 has a unique solution
    • the columns of A are linearly independent
    • A has a pivot in every column
  • the following are equivalent for a (m × n) matrix A:
    • Ax = b has a solution for any choice of b
    • the columns of A span ℝᵐ
    • A has a pivot in every row
  • for a (m × n) matrix:
    • if m < n, then its columns are not linearly independent
    • if n < m, then its columns do not span ℝᵐ

Counterexample you should know

  • a consistent system of linear equations with more equations than unknowns
  • an inconsistent linear of linear equations with more unknowns than equations
  • two distinct echelon forms that are row-equivalent, but have the same entries in every pivot position
  • two different sized sets of vectors which have the same span
  • a set of linearly dependent vectors in which one cannot be written as a linear combination of other others
  • a set of linearly dependent vectors such that every proper subset is linear independent
  • a linear transformation which changes the length of some but not all vectors
  • a linear transformation which changes the direction of some but not all vectors