Simulation Of Stochastic Networks Research Articles

Digital filtering of stochastic-network decision-node noise

Digital filtering techniques are utilized to counteract serial correlation in random-number sequences generated during discrete-event stochastic-network simulations. An algorithm is described which monitors departures from Bernoulli decision-node branch probabilities and compensates for observed serial correlation. Relationships between branch probability and system performance are then explored.

Parallel-Channel, Dual-Resource Constrained Queueing Systems with Heterogeneous Resources

Abstract This paper presents an analysis of a class of multi-resource constrained queueing systems. The systems are “job-shop” queueing systems where both machines and laborers are limiting resources. The model consists of M parallel channels, each containing a single machine. Each service channel has its own queue in which an FCFS discipline is enforced. Arrivals to the system are Poisson and are randomly assigned to a specific service channel upon arrival. The labor force consists of N laborers (N < M) who are not equally skilled; thus, the processing time (exponential) is dependent upon the laborer and machine center utilized. Two different means of allocating labor to competing jobs are considered: (1) a first-in-first-out rule, and (2) a maximum laborer efficiency rule. The emphasis of the analysis is on the effects of the pattern of heterogeneity of the labor force and method of labor allocation. GERTS QR, a stochastic network simulation model, is utilized as a vehicle of analysis.

GERTS QR: A Model for Multi-Resource Constrained Queueing Systems Part II: An Analysis of Parallel-Channel, Dual-Resource Constrained Queueing Systems with Homogeneous Resources

Abstract This paper presents an application of a new model for analysis of a special class of multi resource constrained queueing systems. The systems are “job-shop” systems where both machines and labor are limiting resources. The job-shop model has M parallel channels consisting of a work center containing a single machine at which a single processing operation is performed. Each channel has its own queue in which a strict FCFS discipline is enforced; items are not allowed to switch queues or balk. Arrivals to the system are Poisson and are assigned to a specific service channel at arrival time. The labor force consists of N laborers where N < M; it is assumed that all laborers are equally skilled on a given machine (work center). When jobs compete for labor, the labor is assigned on a first-in-first-out basis. An extended stochastic network simulation model and program, GERTS QR, is utilized to model and analyze the systems.

More on Conditioned Sampling in the Simulation of Stochastic Networks

The technique of conditioned sampling has been shown to improve simulation efficiency in the estimation of stochastic activity network duration c.d.f.'s. This paper describes a.method for generalizing the conditioned sampling approach from its current use of product-form estimators to the use of product/convolution-form estimators. Estimators of the latter type are constructed and demonstrated to require fewer samples per realization (hence increased estimation accuracy) in almost all networks. An algorithm for estimator construction is presented and proven to apply to any given activity network. It is also shown that the derived product/convolution-form estimators may require a precedence structure within the sampling sequence which creates their corresponding realizations.