Page 287 - Applied Probability
P. 287

12. Models of Recombination
                                                                                            275
                                   If G(x) is chosen to make the delayed renewal process stationary, then
                                   show that
                                                                  x
                                                  x

                                                                   U(x − y)dG(y),
                                                      = G(x)+
                                                  µ
                                                                 0
                                                                           "  ∞  −λx     (12.21)
                                   where µ is the mean of F(x). If dH(λ)=     e   dH(x) denotes
                                                                  %
                                                                            0
                                   the Laplace transform of the distribution function H(x) defined on
                                   (0, ∞), also verify the identity
                                                                  1 − dF(λ)
                                                                      %
                                                       dG(λ)  =            .
                                                       %
                                                                     µÐ
                                   Finally, prove that the Laplace transform of the density  1 [1 − F(x)]
                                                                                     µ
                                   matches dG(λ).
                                           %
                                 4. Show that Felsenstein’s [9] map function
                                                             1  e 2(2−γ)d  − 1
                                                      θ  =                               (12.22)
                                                             2 e 2(2−γ)d  − γ +1
                                   arises from a stationary renewal model when 0 ≤ γ ≤ 2. Kosambi’s
                                   map function is the special case γ = 0. Why does (12.22) fail to give
                                   a legal map function when γ> 2? Note that at γ = 2 we define
                                   θ =   d  by l’H opital’s rule.
                                       2d+1
                                 5. Continuing Problem 4, prove that Felsenstein’s map function has in-
                                   verse

                                                           1          1 − 2θ
                                                 d =            ln              .
                                                        2(γ − 2)   1 − 2(γ − 1)θ
                                 6. The Carter and Falconer [4] map function has inverse
                                                            1   −1           −1
                                                 −1
                                               M   (θ)  =    [tan  (2θ) + tanh  (2θ)].
                                                            4
                                   Prove that the map function satisfies the differential equation
                                                                         4

                                                      M (d)   =  1 − 16M (d)
                                   with initial condition M(0) = 0. Deduce from these facts that M(d)
                                   arises from a stationary renewal model.
                                                                      2πi
                                 7. Fix a positive integer m, and let w m = e m be the principal mth root
                                   of unity. For each integer j, define the segmental function m α j (x)
                                   of x to be the finite Fourier transform
                                                                  m−1
                                                                1  	   xw k  −jk
                                                   m α j (x)  =       e  m w m  .
                                                                m
                                                                  k=0
                                   These functions generalize the hyperbolic trig functions cosh(x) and
                                   sinh(x). Prove the following assertions:
   282   283   284   285   286   287   288   289   290   291   292