Voratas Kachitvichyanukul

Binomial random variate generation

Existing binomial random-variate generators are surveyed, and a new generator designed for moderate and large means is developed. The new algorithm, BTPE, has fixed memory requirements and is faster than other such algorithms, both when single, or when many variates are needed.

Computer generation of hypergeometric random variates

The paper presents an exact, uniformly fast algorithm for generating random variates from the hypergeometric distribution. The overall algorithm framework is acceptance/ rejection and is implemented via composition. Three subdensities are used, one is uniform and the other two are exponential. The algorithm is compared with algorithms based on sampling without replacement, inversion, and aliasing. A comprehensive survey of existing algorithms is also given.

Noninverse correlation induction: guidelines for algorithm development

Abstract We propose guidelines for future development of random-variate generators that are capable of inducing statistical dependence between simulation replications without incurring the computational burden of numerically inverting the distribution function. Two examples are given: an exponential generator using the inverse transformation and a generic acceptance—rejection generator based on an existing beta generator. A driver program and illustrative Monte Carlo results are discussed.

Correlation induction without the inverse transformation

Inducing correlation between estimators is a common way to try to reduce variance in simulation experiments. To induce the correlation between estimators, random variates are generated as functions of the same random-number streams. Although the optimal correlation induction occurs with the inverse transformation. The inverse can be quite slow compared to other methods for generating random variates. We discuss an approach for generating random variates quickly while still obtaining substantial correlation induction.

Algorithm 668: H2PEC: sampling from the hypergeometric distribution

Let M be the mode of the hypergeometric distribution, which is defined as the integer portion of {(k + l)(nI + l)/(nl + n2 + 2)). The inverse transformation method is used for M max(O, k n2) < 10, and algorithm HBPE is used for M max(O, K n2) I 10. The overall algorithm framework for HBPE is acceptance/rejection and is implemented via composition. Three subdensities are used: uniform for the body of the distribution and an exponential for each tail.

Fast poisson and binomial algorithms for correlationinduction * *This research is partially supported by the Office of Naval Research contract N00014-7942-0832 through Purdue University

Traditionally, exactness, numerical stability and speed are the three main criteria for evaluating algorithms for random variate generation. However, it is sometimes required that the algorithms provide correlation between generated variates for the purpose of inducing dependence among the output of simulation runs. The inverse transformation, which produces optimal correlation induction, often performs poorly in terms of the first three criteria. Algorithms based on composition, rejection, and special properties which often excel in terms of the first three criteria, tend to scramble the use of random numbers, causing many attempts at common random numbers, antithetic variates and external control variates to fail. The concept of obtaining correlation via algorithms other than the inverse transformation is examined here. To demonstrate feasibility, previously developed algorithms for Poisson and binomial random variate generation are modified to obtain both positive and negative correlation between runs....

ef:

The FORTRAN implementation of an exact, uniformly fast algorithm for generating the binomial, random variables is presented. The algorithm is numerically stable and is faster than other published algorithms. The code uses only standard FORTRAN statements and is portable to most computers; it has been tested on the IBM 370, 3033, 4381, DEC VAX 11/780, SUN 3/50, CDC 6500-6600, ENCORE Multimax, and Apple Mcintosh Plus. A driver program is also included.

Algorithm 678: BTPEC: sampling from the binomial distribution

Abstract The formulation process of a simulation model is a complicated and time consuming task. It is usually performed by a system analyst with knowledge of modeling concepts and simulation languages. In this paper, a natural language modeling environment (NLME) designed to assist system analysts to formulate models is discussed. The strategy is to allow analyst to describe their manufacturing systems in common English and, according to the system descriptions, the computer performs the modeling work. Basically, the process of the proposed natural language modeling environment is to mimic human behavior to interpret system descriptions and is divided into two steps: sentences interpretation and model construction. A description of the system architecture is also given.

A natural language system to aid simulation model formulation

This study examines the feasibility and potential benefits of using simulation to aid designers of large industrial processes. A demon stration of simulations is provided in the basic oxygen furnace shop and subsequent steel-making operations prior to rolling slabs. We explore three design situations: (1) a new processing technology, (2) the removal of bottlenecks in current operations, and (3) sensitivities of processing to equipment failures. Overall shop productivity and the time workloads of crews and indi vidual process operators are examined. Contrasts are made on these criteria between the current mode of operation and the potential design situations stated above through simulation ex periments. Results of these simulation experiments provide the basis for economic and ergonomic justification as well as indica tions for further improvements in the ergonomic facets of design.

A simulation model for ergonomic design of industrial processes

This paper describes various ways in which digital computer simulation can be used by ergonomists in the design of industrial systems. Also shown are recent advances in this methodology which make the technique even more effective and features of the technique which require particular concern. A variety of past simulation studies with human factors considerations are cited for detailed use by practitioners.