WSC 2004

WSC 2004 Final Abstracts

Simulation-Based Scheduling Track

Tuesday 8:30:00 AM 10:00:00 AM
Service Industry Scheduling

Chair: Peter Lendermann (Singapore Institute of Manufacturing Technology)

This paper describes an approach to realtime decision-making for quality of service based scheduling of distributed asynchronous data replication. The proposed approach addresses uncertainty and variability in the quantity of data to replicate over low bandwidth fixed communication links. A dynamic stochastic knapsack is used to model the acceptance policy with dynamic programming optimization employed to perform offline optimization. The obtained optimal values of the input variables are used to build and train a multi-layer neural network. The obtained neural network weights and configuration can be used to perform near optimal accept/reject decisions in real-time. Off-line processing is used to establish the initial acceptance policy and to verify that the system continues to perform near-optimally. The proposed approach is implemented via simulation enabling the evaluation of a variety of scenarios and refinement of the scheduling portion of the model. The preliminary results are very promising.

Coupled Heuristic and Simulation Scheduling in a Highly Variable Environment
Olga Bagatourova and Sudhakar K. Mallya (Bank of America)

The topic of this paper is workforce scheduling in a highly variable environment – check proofing and encoding in a bank – where uncertainties include both volume and arrival pattern of the work. A shop floor simulation model and a heuristic algorithm for workforce scheduling using simula-tion to estimate objective function are presented. The pro-posed approach results in a robust workforce schedule that can be easily adjusted in real time in response to environ-mental changes, such as volume and throughput fluctua-tions. A planning and scheduling framework within which the simulation is implemented is briefly discussed.

Simulation Analysis of Truck Driver Scheduling Rules
Eric C. Ervin (J.B. Hunt Transport, Inc.) and Russell C. Harris (J.B. Hunt Transport, Inc)

2004 brought a landmark event in the changes to regula-tions governing hours of service for truck drivers. This paper describes an effort utilizing modeling and simulation for evaluating the impact of the new 2004 Hours of Service (HOS) rules in scheduling and dispatching one of the largest random over-the-road (OTR) trucking fleets in North America. The model was comprehensive and enterprise-wide in nature, modeling unique order-to-delivery process characteristics for over 120,000 freight lanes and the continuous nature of the driver’s work day. Model results provided quantification of the 2004 HOS impact on fleet utilization, cycle times and customer service. Results of the model were used to guide company strategy related to drivers, customers and operations. With five months of actual business performance collected regarding the new HOS in 2004, a post-mortem analysis has provided insight regarding the quality of simulation model forecasts done in 2003.

A Near Optimal Approach to Quality of Service Data Replication Scheduling
Kevin Adams (Naval Surface Warfare Center Dahlgren Division) and Denis Gracanin and Dušan Teodorovic (Virginia Polytechnic Institute and State University)

Tuesday 10:30:00 AM 12:00:00 PM
Factory Scheduling

Chair: Oliver Rose (Dresden University of Technology)

Monte Carlo techniques have long been used (since Buffon’s experiment to approximate the value of š by tossing a needle onto striped paper) to analyze phenomena which, due to their complexity and/or stochasticity, are beyond the reach of closed-form equations. Basic examples of such studies are estimating the probability that military field communications will remain intact in the face of attack or the number of fish in an irregularly shaped lake. Likewise, scheduling is a necessity for the planning, control, and im-plementation of increasingly large projects in manufacturing, civil construction, military operations, and many other fields. We provide a framework for applying scheduling algorithms based on Monte Carlo simulation, to provide a scheduler, who inevitably confronts numerous uncertainties, an inexpensive and a highly customizable tool that can be utilized in a common spreadsheet environment.

A Heuristic to Determine Equipment Setup Changes Based on Estimated Lot Arrivals in a Semiconductor Fab
Raja Sunkara and Ramesh Rao (National Semiconductor Corp.)

Certain classes of tools used in the semiconductor industry require the tools to be setup differently in order to process different types of products. In cases of large setup times, it is important to minimize the number of setup changes in order to improve the overall equipment utilization. Mini-mizing the number of setup changes needs to be balanced with the restriction on product queue times. Further, it is important that the cycle time of low volume products is not penalized in order to improve equipment utilization. This article presents a heuristic algorithm to determine the setup for each tool in a workstation based on the estimated arrival times of different products at the workstation. The approach described here takes into account the number of tools, their capability, and the expected workload for each setup over a predetermined horizon. The heuristic is independent of the product mix released into the line.

Improving the Performance of Dispatching Rules in Semiconductor Manufacturing by Iterative Simulation
Lars Mönch and Jens Zimmermann (Technical University of Ilmenau)

In this paper, we consider semiconductor manufacturing processes that can be characterized by a diverse product mix, heterogeneous parallel machines, sequence-dependent setup times, a mix of different process types, i.e. single-wafer vs. batch processes, and reentrant process flows. We use dispatching rules that require the estimation of waiting times of the jobs. Based on the lead time iteration concept of Vepsalainen and Morton (1988), we obtain good waiting time estimates by using exponential smoothing techniques. We describe a database-driven architecture that allows for an efficient implementation of the suggested approach. We present results of computational experiments for reference models of semiconductor wafer fabrication facilities. The results demonstrate that the suggested approach leads to high quality solutions.

On Using Monte Carlo Methods for Scheduling
Samarn Chantaravarapan and Ali Gunal (Production Modeling Corporation) and Edward J. Williams (University of Michigan – Dearborn)