← Back to Study Kit

10.018 MSS – Linear Algebra (Wk 8-13) – Side 1

Systems of Linear Equations

  • REF: Leading 1 in each row, zeros below pivots, each pivot right of above
  • RREF: REF + zeros above pivots + pivot cols have only the leading 1
  • Parametric last row \([0\;0\;f(a)\;|\;g(a)]\): No soln if \(f(a)=0, g(a)\neq0\); Unique if \(f(a)\neq0\); Inf. if \(f(a)=0, g(a)=0\)
  • Homogeneous \(A\mathbf{x}=\mathbf{0}\): always consistent; if \(n>m\) ⟹ ∞ solutions

Matrix Operations & Inverses

  • \((AB)^T = B^T A^T\); \((AB)^{-1} = B^{-1}A^{-1}\)
  • 2×2 inverse: \(\frac{1}{ad-bc}\begin{bmatrix}d&-b\\-c&a\end{bmatrix}\)
  • Gauss-Jordan: \([A|I] \to [I|A^{-1}]\)
  • Elementary matrices: Swap (self-inv), Scale \(k\) (inv: \(1/k\)), Add \(kR_j\to R_i\) (inv: subtract)
  • \(A = E_1^{-1}\cdots E_k^{-1}\) if \(E_k\cdots E_1 A = I\)

Determinants & Trace

  • Cofactor: \(C_{ij}=(-1)^{i+j}\det(M_{ij})\)
  • Via row reduction: \(\det = \frac{\text{prod of REF diag}}{(-1)^{\text{swaps}}\times\text{prod of scale constants}}\)
  • \(\det(AB)=\det A\det B\); \(\det(cA)=c^n\det A\); \(\det(A^{-1})=1/\det A\); \(\det(A^T)=\det A\)
  • Triangular \(\det\) = product of diagonal entries
  • \(\text{tr}(A)=\sum a_{ii}\); \(\text{tr}(AB)=\text{tr}(BA)\); \(\text{tr}(A+B)=\text{tr}(A)+\text{tr}(B)\)

Subspaces, Span, Linear Independence

  • Subspace test: contains \(\mathbf{0}\) + closed under \(+\) and scalar \(\times\)
  • LI check: form matrix → row reduce → all columns pivot ⟹ LI
  • \(\text{span}\{v_1,...,v_k\}\) = all linear combos; span is always a subspace

10.018 MSS – Linear Algebra (Wk 8-13) – Side 2

Basis, Rank, Null Space

  • Basis = LI + spans the space; \(\dim\) = # basis vectors
  • row(A) basis: nonzero rows of RREF(A)
  • col(A) basis: pivot columns of ORIGINAL A
  • null(A): solve \(A\mathbf{x}=\mathbf{0}\), parametrise free vars
  • Rank-Nullity: \(\text{rank}(A)+\text{nullity}(A)=n\)
  • Rank = # pivots = dim(row) = dim(col)

Change of Basis

  • \([\mathbf{x}]_E = B[\mathbf{x}]_B\); \([\mathbf{x}]_B = B^{-1}[\mathbf{x}]_E\)
  • Between bases: \([\mathbf{x}]_B = B^{-1}A[\mathbf{x}]_A\)
  • Orthogonal basis coords: \(c_i = \frac{\mathbf{u}\cdot\mathbf{v}_i}{\mathbf{v}_i\cdot\mathbf{v}_i}\)

Eigenvalues & Diagonalisation

  • \(A\mathbf{v}=\lambda\mathbf{v}\); solve \(\det(A-\lambda I)=0\)
  • 2×2: \(\lambda^2 - \text{tr}(A)\lambda + \det(A) = 0\)
  • Eigenspace \(E_\lambda = \text{null}(A-\lambda I)\)
  • alg mult ≥ geo mult ≥ 1
  • Diagonalisable ⟺ geo = alg for all \(\lambda\)
  • \(A=PDP^{-1}\); \(A^k=PD^kP^{-1}\)
  • \(\det(A)=\prod\lambda_i\); \(\text{tr}(A)=\sum\lambda_i\)
  • Markov steady state: eigenvector for \(\lambda=1\), normalise

Linear Transformations

  • \(T\) linear: \(T(\mathbf{u}+\mathbf{v})=T(\mathbf{u})+T(\mathbf{v})\), \(T(c\mathbf{u})=cT(\mathbf{u})\); \(T(\mathbf{0})=\mathbf{0}\)
  • Matrix rep: \(A=[T(\mathbf{e}_1)|\cdots|T(\mathbf{e}_n)]\)
  • From pairs: \([T]=[w_1\cdots w_n][u_1\cdots u_n]^{-1}\)
  • Composite \(S\circ T\): matrix = \(BA\)

Standard Transforms (ℝ²)

  • Rot CCW θ: \(\begin{bmatrix}\cos\theta&-\sin\theta\\\sin\theta&\cos\theta\end{bmatrix}\) CW θ: \(\begin{bmatrix}\cos\theta&\sin\theta\\-\sin\theta&\cos\theta\end{bmatrix}\)
  • Reflect x-axis: \(\begin{bmatrix}1&0\\0&-1\end{bmatrix}\) y-axis: \(\begin{bmatrix}-1&0\\0&1\end{bmatrix}\)
  • Reflect y=x: \(\begin{bmatrix}0&1\\1&0\end{bmatrix}\) y=−x: \(\begin{bmatrix}0&-1\\-1&0\end{bmatrix}\)
  • Reflect line at θ: \(\begin{bmatrix}\cos2\theta&\sin2\theta\\\sin2\theta&-\cos2\theta\end{bmatrix}\)

Projection & Least Squares

  • Proj onto dir a: \(P=\frac{\mathbf{a}\mathbf{a}^T}{\mathbf{a}^T\mathbf{a}}\); \(P^2=P\)
  • Distance: \(\|\mathbf{v}-P\mathbf{v}\|\)
  • Least squares: \(\mathbf{c}=(A^TA)^{-1}A^T\mathbf{b}\)

Common Exam Steps

  • Parametric system: row reduce → check last row condition
  • Inverse: \([A|I]\to[I|A^{-1}]\)
  • Det: row reduce (track ops) or cofactor expand along row/col with most 0s
  • Eigenvalues: \(\det(A-\lambda I)=0\) → \(\text{null}(A-\lambda I)\) for each
  • Diagonalise: check geo=alg; \(P\)=[eigenvectors], \(D\)=diag(eigenvalues)
  • \(A^k\): diagonalise → \(PD^kP^{-1}\)