Search the corpus

    Shift-operator eigenvectors and the Fibonacci subspace

    P-BKF79-7

    Area
    Prelim
    Seen at
    BERKELEY
    Years
    1979
    Status
    draft

    Let \(V\) be the vector space of complex sequences and let \[ S(a_1,a_2,a_3,\ldots)=(a_2,a_3,a_4,\ldots). \]

    1. Find the eigenvectors of \(S\).
    2. Let \(W\) be the subspace of sequences satisfying \[ x_{n+2}=x_{n+1}+x_n. \] Show that \(W\) is two-dimensional and \(S\)-invariant, and give an explicit basis.
    3. Derive an explicit formula for the Fibonacci numbers \(f_1=f_2=1\), \(f_{n+2}=f_{n+1}+f_n\).
    Solution

    1. A nonzero sequence \(a=(a_1,a_2,\ldots)\) is a \(\lambda\)-eigenvectorof \(S\) if and only if\[ a_{n+1}=\lambda a_n \]for every \(n\geq1\).

    The equation \(S(a)=\lambda a\) is the coordinatewise equality \[ (a_2,a_3,\ldots) = (\lambda a_1,\lambda a_2,\ldots), \] which is equivalent to the displayed recurrence.

    2. For every \(\lambda\in\CC\), the \(\lambda\)-eigenspace of \(S\) is\[ \boxed{ \operatorname{span}_{\CC} \{(1,\lambda,\lambda^2,\ldots)\}. } \]

    By step 1, \[ a_n=a_1\lambda^{n-1} \] for all \(n\geq1\). Thus every nonzero \(\lambda\)-eigenvector is a nonzero scalar multiple of \[ (1,\lambda,\lambda^2,\ldots), \] and this sequence is directly checked to satisfy \(S(a)=\lambda a\).

    3. A sequence in \(W\) is uniquely determined by its first twocoordinates.

    Given \(x_1,x_2\in\CC\), the recurrence \[ x_{n+2}=x_{n+1}+x_n \] successively determines \(x_3,x_4,\ldots\). Conversely, the resulting sequence satisfies the recurrence by construction.

    4. The map\[ \Phi:W\longrightarrow\CC^2, \qquad \Phi(x)=(x_1,x_2), \]is a linear isomorphism. In particular,\[ \boxed{\dim_{\CC}W=2}. \]

    The recurrence is homogeneous and linear, so \(W\) is a vector subspace. The map \(\Phi\) is linear, and step 3 says exactly that it is bijective.

    5. The subspace \(W\) is \(S\)-invariant.

    Let \(x\in W\) and put \(y=Sx\), so \(y_n=x_{n+1}\). Then \[ \begin{aligned} y_{n+2} &=x_{n+3}\\ &=x_{n+2}+x_{n+1}\\ &=y_{n+1}+y_n. \end{aligned} \] Hence \(y\in W\).

    6. Set\[ \alpha=\frac{1+\sqrt5}{2}, \qquad \beta=\frac{1-\sqrt5}{2}. \]Then\[ \alpha^2=\alpha+1, \qquad \beta^2=\beta+1. \]

    The numbers \(\alpha\) and \(\beta\) are the two roots of \[ r^2-r-1=0. \]

    7. The sequences\[ u=(1,\alpha,\alpha^2,\ldots), \qquad v=(1,\beta,\beta^2,\ldots) \]belong to \(W\).

    For \(u\), step 6 gives \[ u_{n+2} = \alpha^{n+1} = \alpha^{n-1}\alpha^2 = \alpha^n+\alpha^{n-1} = u_{n+1}+u_n. \] The same calculation with \(\beta\) proves the assertion for \(v\).

    8. The sequences \(u\) and \(v\) form the explicit basis\[ \boxed{ \{(1,\alpha,\alpha^2,\ldots), (1,\beta,\beta^2,\ldots)\} } \]of \(W\).

    By step 2, \(u\) and \(v\) are eigenvectors of \(S\) with distinct eigenvalues \(\alpha\neq\beta\), so they are linearly independent. Step 4 gives \(\dim W=2\), hence they form a basis.

    9. Define\[ F_n \coloneqq \frac{\alpha^n-\beta^n}{\sqrt5}. \]Then \(F_1=F_2=1\).

    Since \[ \alpha-\beta=\sqrt5 \] and \[ \alpha+\beta=1, \] one has \[ F_1=1 \] and \[ F_2 = \frac{(\alpha-\beta)(\alpha+\beta)}{\sqrt5} = 1. \]

    10. The sequence \((F_n)\) satisfies the Fibonacci recurrence\[ F_{n+2}=F_{n+1}+F_n. \]

    By step 6, \[ \alpha^{n+2}=\alpha^{n+1}+\alpha^n \] and \[ \beta^{n+2}=\beta^{n+1}+\beta^n. \] Subtracting and dividing by \(\sqrt5\) gives the claim.

    11. The Fibonacci numbers are\[ \boxed{ f_n = \frac1{\sqrt5} \left[ \left(\frac{1+\sqrt5}{2}\right)^n - \left(\frac{1-\sqrt5}{2}\right)^n \right]. } \]

    Steps <1>9–<1>10 show that \((F_n)\) satisfies the same recurrence and the same initial conditions as \((f_n)\). A second-order recurrence is uniquely determined by its first two terms, so \(f_n=F_n\) for every \(n\geq1\).

    12. Q.E.D.

    Step 2 answers part 1, steps <1>4–<1>8 answer part 2, and step 11 answers part 3.