Batch arrival processor-sharing with application to multi-level processor-sharing scheduling (Q2572914): Difference between revisions

The paper investigates a processor-sharing queue with batch arrivals and Poisson arrival flow. For this queue, the expected conditional response time \(T_{\text{BPS}} (x)\) satisfies the following integral equation \[ T_{\text{BPS}}'(x) = \lambda E[N] \int_0^\infty T_{\text{BPS}}'(y) \overline {F}(x + y)\,dy+ \lambda E[N] \int_0^x T_{\text{BPS}}'(y) \overline {F}(x - y)\,dy + b\overline {F}(x) + 1, \] where \(\lambda \) is the batch arrival rate, \(E[N]\) is the average batch size, \(b + 1 = E[N^2 ]/E[N]\) and \(\overline {F}(x) = 1 - F(x)\) is the complementary distribution of the service time. It is shown that there exists a unique solution to this integral equation. Let \(X\) be a random variable with distribution function \(F(x)\), and let \(\rho \) be the load in the system. The authors provide asymptotical analysis of \(T_{\text{BPS}} (x)\) and prove that \(\lim_{x \to \infty } (T_{\text{BPS}} (x) - \frac{x} {{1 - \rho }}) = \frac{{bE[X](2 - \rho )}} {{2(1 - \rho ^2 )}}\), provided that the service time distribution has a finite mean, the batch size distribution has a finite second moment and \(\rho < 1\). In addition, they give lower and upper bounds for the expected conditional and unconditional response times.