William E. Biles

A gradient—regression search procedure for simulation experimentation

This paper examines a gradient search procedure for simulation experimentation with constrained systems. This procedure combines gradient search with curvilinear regression in moving toward a constrained optimal solution for a system involving n controllable variables. In a direction-determining block, at least n+1 simulation trials are performed around a current base point to establish an improving direction. Then in a step determining block, t simulation trials are performed along the improving direction to establish the most favorable step in moving to the next base point. This sequential block process, in which each block is executed in one input to the computer, is repeated until an approximate solution is found which satisfies all system constraints.

Statistical considerations in simulation on a network of microcomputers

Statistical considerations in simulating complex systems on a hierarchical network of low-cost microcomputers are discussed. Network configurations, allocation of computer tasks to nodes of the network, and assignment of simulation trials are examined. The binary tree and its X-tree variant are found to be especially attractive network configurations in simulation applications with simulation tasks assigned to nodes according to a model-frame/experiment-frame/output-frame trichotomy. Procedures for performing designed simulation experiments, including factor screening experiments, and for conducting variance reduction through manipulation of random number streams are presented.

Two-stage production lines with a single repair crew

SUMMARY This paper describes the analytic modelling of two-stage production lines with storage of workpieces between stages. The effect of stage failure, and repair by a single repair crew, on production rate and intermediate storage level are shown.

Introduction to simulation

Simulation is a valuable tool in the study of complex systems. This paper discusses the fundamental principles of simulation. Its intent is to provide sufficient background in simulation that the potential user would be encouraged to follow through with additional study and soon come abreast of the techniques of simulation.

Kriging metamodeling in multiple-objective simulation optimization

This paper describes the application of Kriging metamodeling in multiple-objective simulation optimization. An Arena-based simulation model of an (s, S) inventory system is utilized to demonstrate the capabilities of Kriging metamodeling as a simulation tool. Response surface methodology and Kriging metamodeling are compared to determine the situations in which one approach might be preferred over the other. The optimization approaches described here have the objective of finding the optimal values of reorder point s and maximum inventory level S so as to minimize the total cost of the inventory system while maximizing customer satisfaction. This paper describes two alternative approaches to utilizing Kriging methodology with multiple-objective optimization in simulation studies.

Animated graphics and computer simulation

This paper discusses the role of animated graphics in computer simulation. Its focus is on low-cost software and hardware, in contrast to the expensive programs and equipment usually available commercially. We see that animated graphics can enhance model development as well as provide a better understanding of the basic mechanisms underlying the system under study.

Optimization of simulation responses in a multicomputing environment

This paper describes the application of experimental design techniques to computer simulation in a multicomputing environment. Three principal areas of experimental design are considered: (1) factor screening experiments; (2) experiments of comparison; and (3) response surface methodology.

Constrained sequential-block search in simulation experimentation

This paper describes the application of sequential-block search techniques to simulation experimentation with constrained systems. Two basically different approaches are examined. One approach combines designed experiments, multiple regression, and mathematical optimization to predict a constrained optimum solution, which is then checked by further experimentation in the region of the predicted solution. A second approach employs a sequential optimum seeking-technique, such as gradient search or sequential simplex search, modified to accommodate constraints. These techniques are illustrated with a simple inventory system modeled with the GASP-II simulation language. A comparison of the effectiveness of these approaches is presented.

Microcomputer-based modeling and simulation

Abstract Modeling and simulation using digital computers is a favorite tool of operations researchers. While it has been used extensively on large computers, simulation on small computers is still in its infancy. This paper takes a look at the state of microcomputer-based simulation and presents a Pascal-based generalized simulation language and statistical support software, Pasmap , which has been developed to facilitate computer simulation on 16-bit microcomputer systems. Pasmap is a compact, portable, event-oriented simulation language which allows discrete-event, continuous, and state-event modeling, as well as specialized statistical analysis. It is a full-fledged attempt to transport Gasp-iv/Simscript-ii.5/Slam -like simulation technology to microcomputer users.

Simulation methodology II

This session focuses on new techniques to assist practitioners of simulation in obtaining desired results efficiently. Many such users are attempting to find optimum performance of a simulated system. In this situation, the problem of selecting a procedure to search for the best choice is a challenging one. Two papers in this session compare alternate strategies for attacking this question. Other users face the task of finding confidence intervals for quantities obtained from simulation experiments. This job is often complicated by statistical dependence of successive observations. The third paper in this session suggests a way around this difficulty by utilizing properties found in many stable stochastic systems.