Search the corpus

    Real matrix with prescribed invariant subspaces and minimal polynomial

    P-BERK80S-01

    Area
    Prelim
    Seen at
    BERKELEY
    Years
    1980
    Status
    draft

    Exhibit a real \(3\times3\) matrix having minimal polynomial \((t^2+1)(t-10)\) which, as a linear transformation of \(\mathbb R^3\), leaves invariant the line \(L\) through \((0,0,0)\) and \((1,1,1)\) and the plane through \((0,0,0)\) perpendicular to \(L\).

    Solution

    Let \[ e=(1,1,1),\qquad u=(1,-1,0),\qquad v=(1,1,-2). \] Then \(L=\operatorname{span}(e)\) and \[ L^\perp=\operatorname{span}(u,v), \] since \(u\cdot e=v\cdot e=0\) and \(u,v\) are linearly independent. Define \(T:\RR^3\to\RR^3\) on the basis \(e,u,v\) by \[ T(e)=10e,\qquad T(u)=v,\qquad T(v)=-u. \]

    1. The line \(L\) and the plane \(L^\perp\) are \(T\)-invariant.

    Because \(T(e)=10e\), the line \(L=\operatorname{span}(e)\) is invariant. Also \(T(u)=v\) and \(T(v)=-u\), so \(T\) maps the spanning set \(\{u,v\}\) of \(L^\perp\) into \(L^\perp\). Hence \(L^\perp\) is invariant.

    2. In the standard basis, \(T\) has matrix\[ \boxed{ A=\frac13 \begin{pmatrix} 11&8&11\\ 12&9&9\\ 7&13&10 \end{pmatrix}.} \]

    Every \(x=(x_1,x_2,x_3)\) has the unique decomposition \[ x=a e+b u+c v, \] where \[ a=\frac{x_1+x_2+x_3}{3},\qquad b=\frac{x_1-x_2}{2},\qquad c=\frac{x_1+x_2-2x_3}{6}. \] Therefore \[ T(x)=10a e+bv-cu. \] Substituting the expressions for \(a,b,c\) gives \(T(x)=Ax\).

    3. The minimal polynomial of \(A\) is \((t^2+1)(t-10)\).

    On \(L\), the restriction of \(T\) is multiplication by \(10\), so its minimal polynomial is \(t-10\). On \(L^\perp\), relative to the basis \((u,v)\), the restriction has matrix \[ \begin{pmatrix}0&-1\\1&0\end{pmatrix}, \] whose square is \(-I\) and which is not a scalar matrix. Hence its minimal polynomial is \(t^2+1\).

    Since \[ \RR^3=L\oplus L^\perp \] and both summands are invariant by step 1, the minimal polynomial of \(T\) is the least common multiple of the minimal polynomials of the two restrictions. The polynomials \(t-10\) and \(t^2+1\) are coprime over \(\RR\), so \[ \mu_A(t)=\operatorname{lcm}(t-10,t^2+1)=(t-10)(t^2+1). \]

    4. Q.E.D.

    Step 2 exhibits the matrix \(A\), step 1 shows that it leaves \(L\) and \(L^\perp\) invariant, and step 3 gives its minimal polynomial.