0

For investigating a queueing model in my research, I need to obtain the probability generating function (PGF), $G(z)$, by solving the following functional equation $$\lambda(z(1-p)+p)G((1-p)z)-(\mu+\lambda p)(1-p)G(z)=(\lambda p-(\mu+\lambda p)(1-p))\pi_0$$ where $\lambda, \mu, p, \pi_0$ are parameters, and $0<p<1$.

Since $G(z)$ is a PGF, it is known that $G(0)=\pi_0$ and $G(1)=1$. Also, by substituting $z=1$, I get $$G(1-p)=\frac{(\lambda p-(\mu+\lambda p)(1-p))\pi_0+(\mu+\lambda p)(1-p)}{\lambda}$$ In addition, by substituting $z=1/(1-p)$, $z=1-p$ and $z=(1-p)^2$, I have $$G(\frac{1}{1-p})=\frac{\lambda(1+p)-(\lambda p-(\mu+\lambda p)(1-p))\pi_0}{(\mu+\lambda p)(1-p)})$$ $$G((1-p)^2)=\frac{(\lambda p-(\mu+\lambda p)(1-p))\pi_0+(\mu+\lambda p)(1-p)G(1-p)}{\lambda}$$ $$G((1-p)^3)=\frac{(\lambda p-(\mu+\lambda p)(1-p))\pi_0+(\mu+\lambda p)(1-p)G(1-p)^2}{\lambda}$$ Since $G(z)$ is a PGF it cannot be a polynomial function. So, I assumed that the function is of form $$G(z)=\frac{az^2+bz+c}{dz^2+ez+f}$$ and tried to calculate the coefficients $a, b, c, d, e, f$ by solving the set of six equations using the six expressions above, i.e., $$\frac{c}{f}=\pi_0$$ $$\frac{a+b+c}{d+e+f}=1$$ $$\frac{a(1-p)^2+b(1-p)+c}{d(1-p)^2+e(1-p)+f}=G(1-p)$$ $$\frac{a(1/(1-p))^2+b(1/(1-p))+c}{d(1/(1-p))^2+e(1/(1-p))+f}=G(1/(1-p))$$ $$\frac{a(1-p)^4+b(1-p)^2+c}{d(1-p)^4+e(1-p)^2+f}=G((1-p)^2)$$ $$\frac{a(1-p)^6+b(1-p)^3+c}{d(1-p)^6+e(1-p)^3+f}=G((1-p)^3)$$

However, there is no possible solution for this set. I will appreciate if some one could help me.

  • Welcome to MSE. Your question is phrased as an isolated problem, without any further information or context. This does not match many users' quality standards, so it may attract downvotes, or be closed. To prevent that, please [edit] the question. This will help you recognize and resolve the issues. Concretely: please provide context, and include your work and thoughts on the problem. These changes can help in formulating more appropriate answers. – José Carlos Santos Aug 31 '23 at 11:13
  • 1
    Denote $q=1-p.$ If you write $G(z)=\sum_{n=0}^{\infty}a_nz^n$ then $$a_n=G(0)\prod_{k=0}^{n-1}\frac{1}{p-(\mu+\lambda p)q^{-k}}.$$ Certainly $G$ is not elementary, in particular is not the quotient of two polynomials. Expanding a rational fraction in a power series is done by writing it in sum of $(1-rz) ^{-k}$, essentially giving $a_n=\sum_{r\in R}P_r(n)r^n$ when $n$ is big enough -Here $R$ is a finite set of complex numbers and $P_r$ is a polynomial. So, no way... – Gérard Letac Sep 03 '23 at 09:38

0 Answers0