Showing papers on "Service system published in 1972"

Multilevel Bulk Service Queues

Abstract: This paper uses the imbedded-Markov-chain method to study bulk service queuing systems in which the server chooses the capacity and service mechanism depending on the number of customers waiting in line at the start of the service. It derives conditions under which the systems studied are ergodic and obtains the stationary state probability generating functions. A bilevel hysteretic bulk service system is studied in detail using numerical methods. A concept of system "scale" is introduced that permits the calculation of operating characteristics of large systems from those of small systems. This property is important because the computations for large systems are much more difficult, if not impossible, to perform. The characteristics studied are the expectation and variance of the queue size, the proportion of time each service level is used, and the transition rate between service levels.

The Application of Multifunction Stochastic Service Systems in Allocating Ambulances to an Urban Area

Abstract: Alternative policies for the allocation and distribution of ambulances are studied for the city of Detroit, Michigan. Data on emergency occurrence and service processes were collected and analyzed, and these data analyses are utilized to model the ambulance system as a multifunction stochastic service system with semi-Markov arrivals and state-dependent server selection. Numerical results are developed predicting the performance of both single function recovery systems and dual function police-ambulance systems under alternative operating policies.

A Problem in Bulk Service Scheduling

Abstract: Given the operating costs, usage function, and planning horizon of a bulk service system, say a common-carrier transportation system, the objective is to determine the number and the starting times of the services so that the total profit is maximized For such problems, this paper studies the existence and computability of optimal schedules It shows that an ϵ-optimal schedule can be found by a discrete optimization on a sufficiently fine grid It then makes more specific assumptions regarding the dependence of usage on the schedule and presents a dynamic-programming algorithm for the discrete optimization Finally, an analytic solution is given for a simple case

The Optimality of Connected Policies for Markovian Systems with Two Types of Service

Abstract: Consider a single station service system with bounded queue capacity and two possible types of service (“fast” and “slow”), operating in discrete time. It is proved that in the search for an optimal policy one need not consider policies hat prescribe one type of service when the system is relatively empty, another type when there are more customers in the system, and the first type again if there are even more customers in the system.

Asymptotic Service System Output, with Application to Multiprogramming

Abstract: In a multiprogramming computer system modeled as a cyclic queue the number of program segments completed at the CPU in a period of time is discussed. It is shown that this number is approximately normally distributed as the time period becomes long. Parameters of the normal distribution are determined. Numerical examples illustrate the results.