Analysis of the \(M^{[X]}/G/1\) queues with second multi-optional service and unreliable server (Q993666)

The paper studies an \(M^X/G/1\) queueing system, where customers arrive in the system according to compound Poisson process with rate \(\lambda\) and batch size \(X\) having a discrete distribution \(\mathsf{P}\{X=i\}=C_i\) with mean \(c=\mathsf{E}X\) and \(\mathsf{E}X^{2}<\infty\). Service of the first type is required for all customers and is assumed to have mean \(\frac{1}{\mu_0}\) and finite variance. As soon as the service of the first type is completed, the customer requires service of the second type. This type of service is multi-optional. That is there are \(m\) options of a secondary service time, and there is probability \(r_k\) that the \(k\)th option will be chosen. Then the probability that a secondary service is not required is \(r_0=1-\sum_{i=1}^kr_k\). The \(k\)th option service is assumed to have the mean \(\frac{1}{\mu_k}\) and finite variance. The server is assumed to be unreliable. It breaks down after an exponentially distributed time that depends on option. In the case of the first type service this exponential service time has the mean \(\frac{1}{\alpha_0}\), and in the case of option \(k\) of secondary service, this exponential distribution has the mean \(\frac{1}{\alpha_k}\). The repair times are generally distributed and have second moments, and their corresponding distribution functions are denoted \(G_k(x)\), \(0\leq k\leq m\). By using the supplementary variable method the paper derived performance characteristics for transient and steady-state behavior of the system.