Scribd Logo
Upload

Welcome to Scribd!

  • 18 views

    Tang Port

    Uploaded by

    Fatima Luyando
    This document summarizes the proof of finding the maximum of a function f(x) defined as: f(x) = (hx,Ri - Rf)/(x^TCx)^(1/2). It defines three concepts needed for the proof, shows that f(x) agrees with the function g(x) defined as g(x) = hx,R̄i/(x^TCx)^(1/2) where R̄ = R - Rf*e, and proves that g(x) is constant on rays. It then takes the derivative of g(x) and sets it equal to 0 to find the maximum, which is given by x

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Tang Port

    Uploaded by

    Fatima Luyando
    18 views3 pages
    This document summarizes the proof of finding the maximum of a function f(x) defined as: f(x) = (hx,Ri - Rf)/(x^TCx)^(1/2). It defines three concepts needed for the proof, shows that f(x) agrees with the function g(x) defined as g(x) = hx,R̄i/(x^TCx)^(1/2) where R̄ = R - Rf*e, and proves that g(x) is constant on rays. It then takes the derivative of g(x) and sets it equal to 0 to find the maximum, which is given by x

    Original Title

    TangPort

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes the proof of finding the maximum of a function f(x) defined as: f(x) = (hx,Ri - Rf)/(x^TCx)^(1/2). It defines three concepts needed for the proof, shows that f(x) agrees with the function g(x) defined as g(x) = hx,R̄i/(x^TCx)^(1/2) where R̄ = R - Rf*e, and proves that g(x) is constant on rays. It then takes the derivative of g(x) and sets it equal to 0 to find the maximum, which is given by x

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    18 views3 pages

    Tang Port

    Uploaded by

    Fatima Luyando
    This document summarizes the proof of finding the maximum of a function f(x) defined as: f(x) = (hx,Ri - Rf)/(x^TCx)^(1/2). It defines three concepts needed for the proof, shows that f(x) agrees with the function g(x) defined as g(x) = hx,R̄i/(x^TCx)^(1/2) where R̄ = R - Rf*e, and proves that g(x) is constant on rays. It then takes the derivative of g(x) and sets it equal to 0 to find the maximum, which is given by x

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 3

    Tangent Portfolio

    From: Mathematical Portfolio Theory


    Joshi and Paterson
    March 30, 2020

    1 Background concepts
    This section describes three concepts needed for the proof.

    n
    PFirst we define that functions f and
    P g (functions in R ) agree on the set
    { xi = 1} iff (x) = g(x) for any x { xi = 1}

    Secondly, we say that when a function g is constant on rays ⇒ g(λx) = g(x)


    ∀λ>0

    Lastly, if we consider a function g that is constant on rays and we find the


    maximum (y) of g(x) ⇒ g(y) ≥ g(z) ∀z and g(λy) = g(y) ∀ λ > 0.
    P
    Using all these concepts, if we consider x { xi = 1} ⇒ f (x) = g(x)
    If we now consider z = λy ⇒ g(z) ≥ g(x) ∀x and if we also consider λ = P1
    P yi
    then z { xi = 1} therefore f (z) ≥ f (x) so z is also the maximum of f.

    2 Proof
    We need to find the maximum of this function:
    hx,Ri−Rf
    f (x) = 1
    (xT Cx) 2

    We will define the vector eT = (1, 1, ..., 1)


    P
    Therefore, if x { xi = 1} ⇒ hx, ei = 1 and we can say that f (x) agrees
    hx,Ri−Rf hx,ei
    with g(x) defined as: g(x) = T
    1 = hx,
    T
    R̄i
    1
    (x Cx) 2 (x Cx) 2

    Where, R̄ = R − Rf e

    Now we need to check if g is constant on Rays.

    1
    hλx,R̄i
    g(λx) = 1
    ((λx)T C(λx)) 2
    We will need to evaluate first, hλx, R̄i
    It can be seen that hλx, R̄i = λhx, R̄i
      21 1
    T
    Then we evaluate the denominator and find that (λx) C (λx) = λ xT Cx 2

    Se we can conclude that g is constant on rays.

    Now we will find the maximum of g(x).

    We can write g(x) = k(x)l(x) where,

    k(x) = hx, R̄i


    − 1
    l(x) = xT Cx 2

    We need to get the derivative of a function in Rn , so we need to get all the


    partial derivatives.
    ∂g(x) ∂k(x) ∂l(x)
    ∂xi = ∂xi l(x) + ∂xi k(x)

    We find that:
    ∂k(x)
    ∂xi = R̄i

    And

    ∂l(x) − 3 ∂ (xT Cx) − 3


    ∂xi = − 12 xT Cx 2 ∂xi = − 12 xT Cx 2 2 (Cx)i

    So we have,

    ∂g(x) − 1 h hx,R̄i(Cx)i
    i
    ∂xi = xT Cx 2 R̄i − xT Cx

    ∂g(x)
    We need ∂xi = 0 ∀i, which is equivalen to:

    hx,R̄i(Cx)i
    R̄i = xT Cx
    ∀i

    And:

    R̄ = α(x)Cx

    Where,

    2
    hx,R̄i
    α(x) = xT Cx

    The solution is given by x = C −1 R̄ which is the maximum of g(x). Remember


    that g(x) is constant
    P on rays, so y = λx is the maximum. Also f (y) agrees with
    g(y) since λ = xi .

    So, the tangent portfolio will be given by:

    P1
    
    y= xi x

    You might also like