Using Discrete-Event Simulation to Create Flexibility in APAC Supply Chain Management

Tiger, Andrew A

Abstract

While discrete-event simulations have been used in a number of situations to model projects and events, this paper describes how the methodology may be used in a supply chain management context to assist in decisions where time and flexibility is essential. The model was developed to assist a multi-billion dollar company in understanding the impact of material flow from the US to the Asia-Pacific region (APAC) and to allow flexibility in the supply chain. More specifically, the DSS quantified the impact of shipping directly from US sources of supply to APAC customers versus shipping through consolidation (existing and proposed) locations using cycle time, throughput, work-in-process (WIP) and shipped containers as performance measures. The DSS is organizationally structured through a partnership between academia and industry where the academician maintains the model and industry maintains and supplies the data. This unique partnership provides industry with stable, long-term modeling expertise and increased productivity. Academia benefits from the opportunity by having access to practical global supply chain management issues.

Keywords : decision support system, flexibility, simulation, supply chain management

Introduction

Supply chains encompass the flow of material, production and information from the basic raw materials through delivery to the final customer. Flexible supply chains exhibit characteristics of flexible manufacturing the ability to reduce costs while adapting rapidly to changes in consumer demand. Manufacturing system flexibility often must be designed into manufacturing equipment, thus making the equipment able to quickly change from one product to another. The similarity in supply chain management (SCM) is the ability to change the network between supplier and customer when system changes create the need. Although similar, one significant difference exists. In manufacturing, equipment design requires significant planning time. In SCM, network options are often readily available. This means that optimal network option selection and timing are the critical issues for SCM flexibility.

To address these issues, spreadsheets are often the analytical tool of choice for studying supply chains because of their affordable power and ease-of-use. Unfortunately, the time-based nature of supply chain critical performance measures (cycle time, throughput, and WIP) are not easily reproduced in static spreadsheets. A viable alternative to these problems is use of discrete event simulations (DES), a time-based modeling tool, which allows calculation of time-based statistics. just as important, simulation code and animation provide an understandable representation usually accepted by non-modelers. The DBS-based DSS's purpose is to provide useful knowledge to enable flexible supply chains to quickly adapt to changes in the system.

Several reasons likely account for the slow adoption of simulation in solving SCM process problems. Primarily, simulations are typically used for static, project solutions rather than for process ones. In a project, the output is often a point solution, a one-time, point in time answer; whereas, a process is a re-usable methodology that does not die at the end of a project. This dynamic nature of processes has likely inhibited the use of simulations as the basis of a decision support system (DSS). Other specific difficulties in using simulation for SCM processes solutions include:

1. The lack of readily available data when system changes occur.

2. The modeling tool/software language requires commitment, is expensive, and rapidly obsolete.

3. The lack of modeling expertise.

4. Poor modeling design that prevents adaptation to a changing system.

5. Complexity of many supply chains.

Each of these factors that inhibit the use of simulation in processes, however, has a solution:

1. Data acquisition may be rapid. Technologies exist to generate and automatically integrate massive amounts of information into the DSS.

2. Software languages are continually being improved and competition has significantly created affordable simulation solutions. Additionally, outsourcing of the simulation modeling may reduce the commitment to rapidly obsolete software.

3. Creating a partnership between academia and industry where academia maintains the model and industry maintains the data. The academician maintains the model and the company provides the data when new situations arise that require analysis. An advantage of this partnership is model viability. One primary reason a simulation model is shelved is that the original developer vacates his position. In today's market, professionals are constantly switching career paths. Since a professor usually remains a professor, fewer opportunities exist for the model to be shelved. A benefit of this partnership is maintaining a link between academia and industry. Through these partnerships, the academician has the opportunity to be involved with practical business problems, which may provide a greater competitive advantage in recruiting quality students.

4. DSS design rules must be followed by creating scalable models and separating data from the model. Models must be planned to expect increased complexity because of larger or more detailed requirements. This is discussed in detail in the section on DSS.

The purpose of this paper is to illustrate how a discrete-event simulation may be used to provide solutions for the flexible supply chain problems of timing and selection, using a real life example. In this paper, the usefulness of the DSS is not for daily changes, but for periodic changes in the supply chain design due to fundamental shifts in the system. Examples include new product introductions, new technology in production, a changing customer base due to winning a long term contract, political changes, new laws, and so on. Consequently, the DES model is not run everyday to decide that day's (or near future) operations. Rather, it is used to modify the source of supply, change the transit mode, negotiate rates with carriers, investigate using 3PL's for consolidation of product, etc. Although not run on a daily basis, when the need does arise, an analysis is needed promptly. Without a living model, companies typically use large amounts of human effort to fight these fires. Consequently, the DSS-based analysis' value is equally as attributable to its promptness as to its quality.

Literature Review

DES is a venerable and well-defined methodology of operations research and many excellent explanatory texts exist (Law and Kelton, 2000; Pritsker, 1995; Winston, 2000). The methodology is particularly useful in evaluating interdependencies among random effects that may cause a serious degradation in performance even though the average performance characteristics of the system appear to be acceptable (Shapiro, 2001). As such, DES has been used to study flexibility in manufacturing systems (Albino et al., 1999; Caprihan, 1997; Gupta et al., 1992; Garg et al., 2001; Borenstein, 2000; Pflughoeft, 1999; Nandkeolyar et al., 1992). However, simulation is primarily used to demonstrate flexibility of a design parameter, i.e. routings, polices, and equipment design. These types of simulation models are termed "throw away models" because they are seldom used after designs are finalized (Thompson, 1994). The simulation itself was not the vehicle of flexibility, which is the thesis of this paper. That is, flexibility in supply chain management is achieved through building a decision support process around a simulation model.

Simulation models are intuitive to understand, which is an important reason for their longtime and continuing application to supply chain problems (An et al., 1994; Ferguson, 1998; Chen, 1999). Similarly with manufacturing simulations, supply chain simulations are also primarily of the throw away type and only provide a point solution. In contrast, DES may also serve as a basis for a DSS process. However, a debate has arisen on whether supply chain simulation models are adequate for the basis of decision support systems. Shapiro (2000) cites two "serious deficiencies": The first is the time and effort required. The second is a simulation model fails to provide insights into how a system can be optimized. Neither deficiency is insurmountable, however.

Companies requiring the development and use of complex supply chain simulation models may derive greater efficiencies and use of resources through outsourcing the task to experts as actualized in the example presented in this paper. As for the second deficiency, optimization techniques are inadequate for describing or solving many systems. For those cases, experimentation with simulation is a worthy alternative because the strength of simulations is in their ability to model almost any system, regardless of its complexity. Additionally, in practice, expertise of optimization problems is even more difficult to find than expertise of simulation as the mechanics of simulation models are easier than optimization models to understand because of the animation and modeling constructs that simulation provides. Companies needing analysis have more options with simulation. Finally, the application of operations research within an organization is evolutionary. Companies at the initial stages of evolution should begin with less complex solutions, rather than jumping to the most advanced solutions. Otherwise, failure is likely, and operations research is cast aside.

A current trend of using simulation as an instrument of flexibility is real-time simulation (Wu and Wysk, 1989; Erickson et al., 1987; Harmonosky and Barrick, 1988; Harmonosky, 1990; Cho and Wysk, 1993; Smith et al. 1994; Jones et al., 1995; Peters et al., 1995). Real-time simulation is the use of simulation technology for real-time operational control within manufacturing systems. Real-time simulation emulates the control logic and mimics the behavior of the manufacturing system for short periods of time. Its objective is to provide advanced planning and scheduling capability to aid in capacity planning, sequencing, predicting leadtimes and duedate quoting. Simply put, real-time simulation is a DSS process used on a daily basis with the level of detail is at the shop floor. Users are within the walls of the plant and less emphasis is on representing the stochastic nature of the system. In contrast, the use of simulation as a process is expected to be monthly/quarterly. If daily use of the model was expected, then the academic/ industry partnership would be of little value. Also, the stochastic nature of the system does play a role in the analysis, specifically in demand orders and transit times. Considering the previous use of these real-time simulations in daily operations, the potential benefits of using simulation in a SCM DSS system to facilitate order and transit time flexibility holds promise.

The next section provides an example of how a DES is used to create greater flexibility in the supply chain management of a major firm. The importance of this research has several facets. First, it demonstrates the continuing use of simulation to study supply chain issues, specifically in APAC. second, it demonstrates that simulation can be the basis of a DSS process instead of a "throw-away" model. Finally, it demonstrates that OR can be structurally organized by using an academic/industry partnership that improves the viability of the process.

The DSS for Domestic Source of Supply Material Row to APAC

Problem Description

The company used in this example is a multi-billion dollar technology-based company with over half its sales derived from outside the United States, predominately in the Asia-Pacific (APAC) region. Known as a company of new products, the diversified product line gives the company the ability to meet its customers where they are, whether the customer is in a developing country or a state-of-the-art technological one. However, for this company sources of supply are constantly in a state of fluctuation and its APAC customers experience spiky, highly-variable demand growth patterns. Expansion, socio-economic conditions, and politics make it difficult to forecast consumer demand. Additionally, as a world-wide company, domestic and international supply chains often overlap to take advantage of scale. When redesigning one, it impacts the other. This dependency and cascading effect makes tracking difficult. Finally, the impacts of supply chain design on the performance values of cycle time, throughput, and WIP inventory are difficult to calculate for non-existing systems. Variability and interactions significantly impact these values. Consequently, effective APAC supply chain management offers immense potential savings in improved inventory management.

Not surprisingly, the company's international logistics management group is full of experienced professionals quite knowledgeable of the mechanics of establishing and maintaining operations outside the US. Measurement and incremental improvements in efficiency of existing systems is not the challenge. Rather, investigating new options due to fundamental changes both in supply and demand are the challenge. To meet the challenge, a partnership with academia was formed to create a decision support system (DSS) using discrete-event simulation (DES) that would add flexibility when operating and designing its supply chains in Asia-Pacific (APAC) operations. As a bonus, the DBS-based DSS would improve productivity and free people to perform tasks where creativity is needed.

The system to be modeled is the company's current flow of material from US-based sources of supply (SOS) to APAC. From all parts of the US, more than twenty-five SOS export over 100 million tons of products annually to APAC customers in 20+ countries. The primary SCM decision at hand is whether to ship directly from a SOS to an APAC customer or to ship through a US-based consolidation point in an attempt to reduce transit costs by achieving economical load factors (Figure 1). Provided a full load is ready, shipping directly is always the preferred route both in terms of cost and transit time. However, if not enough orders have generated a full load, the company will ship the current orders to a consolidation point. Since other SOS are doing the same, an APAC customer-specific full load is generated at the US-based consolidation point.

The decision to ship directly or through a consolidation point is complex for several reasons:

1. The consequence of waiting for a full load: Although shipping directly is always the cheapest and fastest, waiting for a full load before shipping increases cycle time and WIP levels.

2. Demand variability: For many reasons, APAC customers generate orders at different volumes and with different degrees of variability, which may substantially increase wait time.

3. Product characteristic: The definition of a 'full-load' is unique due to the many different products the company produces. Some loads 'cube-out' before 'weighing-out' if the product is relatively light.

4. Delivery time components: Delivery times are affected not only of by transit time, but also by carrier/lane availability and border-crossing times. Some APAC countries are trade friendly while others are not.

5. Overlapping supply chains: Although shipping direct is beneficial, shipping through the US-based consolidation point has the advantage of overlapping US/Out of US supply chains because the consolidation point is also a consolidation point for domestic and non-APAC Out of US customers. Therefore, full loads from the SOS to the consolidation point occur often.

Aside from these complexities, the following are questions that the DES logic system must address:

1. Using the existing network, how long should a US-based SOS wait for a full load (versus shipping through the consolidation point)?

2. Should an APAC-based consolidation point exist? If so, where? and with what impact on the domestic supply chain?

3. Can the critical components of the supply chain be determined and used to negotiate with carriers/ logistics companies to minimize shipping costs?

Each of these questions may be answered through a simulation model that allows flexibility when conditions change, despite the complex variables required in the supply chain decisions presented here. specifics of the model developed and the software used are discussed next.

Modeling Issues

Discrete-event simulation was the technique used to model the flexible supply chain timing and load decisions. Software used to create the DES was SimulS. SimulS is a low-priced ($1,000) DES package that offers considerable capability and ease-of-use. Its programming language, Visual Logic (VL), and global variable spreadsheets are invaluable when modeling supply chains.

Initialization Logic Using SmulS's Visual Logic and Global Variable Spreadsheets

For DBS-models to become part of a process, data must be clearly separate from the model logic. Otherwise, subsequent analyses require modeling changes instead of only data changes. The more effort put in designing the separation of data and model, the more likely the model will be appropriately specified and suitable for use in subsequent analysis. In this project, all modeling parameter values were read from a MS-Excel data file. The company generated the MS-Excel file from database systems. When designing the data file, non-existing network possibilities were included to generalize the model. For example, some APAC customers had no history of ordering from some SOS. However, that possibility (and dummy values) was still modeled to (1) provide future possibilities and (2) allow the use of looping logic when initializing the model.

The software's internal programming language resembles many other application-level languages and provides complex logic and general extensibility. In this model, VL was used for initializing the system of work center parameters (Note: a work center is a SimulS construct that performs work on an entity). For example, the definition of a full load was unique and SOS-customer specific. That is, each SOS-customer combination had uniquely defined full loads. By using clever naming conventions of the modeling work centers and using VL looping logic, work center full load values were initialized for each combination. Without this capability, a work center representing each combination would have to be modified manually, likely 200+ potential modifications.

Model parameter data inputted from MS-Excel was imported to global variable arrays defined as 'Information Store Spreadsheets' in SimulS. These spreadsheets can be readily accessed using looping logic and/or indices using VL. Within the model, each day's new orders attributes are assigned by using the spreadsheet and VL looping logic.

Time Periods and Entities

The simulation time periods were defined as days, primarily because customer orders occurred on a daily basis. The SOS operated on 5, 6, or 7 day workweeks and calendar logic represented this accordingly. A simulation entity was 2000 lbs. Several 'scaling factors' were evaluated with the tradeoff of accuracy versus CPU time. Larger scaling factors (i.e., 5,000 lbs., 10,000 lbs.) ran much quicker, but failed to adequately track the number of shipments ( see section on Validation). Smaller scales (1,000 lbs.) required lengthy CPU time (over a day) to run a one-year trial. A 2000 Ib.scale trial ran in 1-2 hours on a Pentium III 1.0 GHz machine.

Simul8 has 'high-volume' capability designed to be used for Fast Moving Consumer Goods (FMCG) applications. It offers reductions in simulation time by reducing the number of entities. Unfortunately, the 'high-volume' capability was incompatible with the complex routing and collecting rules required by the system.

Modeling Waiting for a Full Load Using SmulS's Shelf Life

The primary decision logic was keeping track of how long it took for filled orders to generate a full load. If a full load occurred before some predetermined time period, direct shipping occurred. Otherwise, the filled orders were shipped to the consolidation point. SimulS's "Shelf Life" construct modeled this logic very neatly without having to use extra code. In SimulS, queues can have a shelf life, which is the maximum amount of time that an entity waits in the queue before being removed. In the model, each SOS had a queue of filled orders. Filled orders that waited until the shelf-life value was reached were immediately removed from the queue and transferred to other logic that moved the entities to the consolidation point.

Modeling Transit Times Using SmulS

Tracking filled-orders moving through the supply chain required modeling both order volume and the number of shipments. A shipment is composed of multiple orders and is a dynamic value. Visual logic was used to dynamically determine how many orders were assigned to a shipment. Once assigned, transit times were modeled using Simul8 queue's minimum wait time construct. Similar but opposite to shelf life, minimum wait time is the amount of time that entities must wait before they can leave the queue.

Validation

Validating the model required reproducing the previous year's actual system performance measures in terms of volume (lbs) and shipments (count). Each location (SOS, consolidation point, and APAC customer) was compared by running actual shipment history through the model. Student t-tests for each location's volume and count showed the model statistically represented the actual system. Once validated, the historical shipments were fit to probability distributions. Many were classically Poisson; however, some of the low volume SOS-customer combinations were fit with empirical discrete probability distributions. A warm-up period of three months (ninety days) was used and adequately provided a stabilized time-in-system (Figure 2).

A Sample Analysis : Determining Optimal Wait Time for Direct Shipments

The initial analysis quantified the effect of waiting for a full load on cycle time and WIP. As mentioned, fully loaded, direct shipments between a US-based SOS and an APAC customer were always the fastest and cheapest. Also, full loads have the benefit of reducing the number of shipped containers through the consolidation location. However, waiting for a full-load does add WIP to the system and increases cycle time. Prior to the analysis, the assumed cost tradeoff was shipping costs versus inventory costs.

Figures 3 and 4 show the results. In all charts, the horizontal axis is the "Maximum Days Waiting for a Full Load". For all SOS/customer combinations, an across-the-board value was chosen for the first ten runs. In the last run, an intelligent policy was used that was specific to each SOS/Customer combination. In the policy, reasonable (50% chance) time-to-generate-a-full-load estimates were calculated using a Normal approximation of the Poisson demands. Provided these demand estimates were less than a week, direct shipping was given an opportunity. If the estimates were longer, no direct shipping was used, and the SOS shipped immediately to the consolidation location.

For the first ten runs, Figure 3 shows that WIP increased as SOS waited for a full loadmaterial waiting on a dock adds both time and inventory to the system. However, the expected benefit was assumed to be fewer shipments. Figure 4 shows a reduction in shipments from the consolidation location to APAC customers; however, direct shipments increased. From history and the model, direct shipping loading efficiencies are slightly less than through consolidation. As direct volume increases, direct container counts increase faster than consolidated container counts decrease. Consequently, total shipments gradually increase, and no advantage exists for waiting for a full load.

Using a WIP cost of $1/1b, for a 100 million pound volume, the annual WIP savings equals $273,973.

Figure 4 shows the Smart policy reduced the annual container shipments to 4,600 from the Four-Day Max Wait policy of 5,000, a reduction of 400 containers. Assuming the cost to ship a container is $500, and that the cost of shipping is only a function of the number of containers, irrespective of the loading efficiency, the annual shipping cost savings realized in this example would be $200,000. With both WIP levels and total shipments reduced in this example, the total annual cost savings is $473,973. Although consolidated shipments increased, the impact of double handling was more than offset by the reduction in the total containers shipped .and the reduction in waiting for a full load time.

Project/Process Management

The DSS was developed in three months. This included an initial meeting, model building, validation, and an initial analysis. During the initial meeting, the process objective was established and emphasized the development of a long-term methodology for addressing SCM questions rather than on answering a specific problem. The long-term approach minimized having to rush to provide an answer often seen in simulation projects. secondly, a long-term outlook ensured that cost savings and beneficial results did not have to occur on the initial analysis. Simulation answers often are similar to intuition/guesses. In these instances, the simulation model has little effect on the cost savings. Short-term thinking may conclude that simulation should be abandoned. However, a situation will eventually arise where the model offers a counter-intuitive solution. A long- term outlook ensures these results can occur.

Additionally, the benefit of simulation (and operations research methodology in general) is often seen in increased productivity. Provided enough people and resources are available, most SCM problems can be reasonably solved. Unfortunately, this fire-fighting methodology requires having enough people to attack the problem on an as-needed basis. With corporate down-sizing, firefighting is not a reasonable approach. An operations research-based DSS replaces people effort; thus, productivity is increased. For this DSS, a second analysis was requested six weeks after the initial analysis. The initial analysis was produced in three-months, and the subsequent analysis took three days because of the establishment of the process.

Conclusions

In this paper, flexibility is defined as the distillation of massive amounts of information into useful knowledge to enable supply chains to adapt quickly to changes in the system such as consumer tastes, new technologies, or global political changes such as acts of terrorism and/or increased border security. The use of a decision support system based on discrete-event simulation for modeling material flow from US-based sources of supply to APAC-based customers was demonstrated. The simulation modeling software was SimulS, a full-featured, low-cost solution that adequately represented the supply chain of study.

Flexibility requires answers rapidly. Although supply chain design changes are not generally daily decisions, prompt answers are valuable when a design change is to be considered. To ensure this type of flexibility a partnership was established between academia and industry where the academician maintained the model, and the company maintained and provided the data. Not only does this provide flexibility, but it also reduces economic risk by removing the requirement that the company add modeling expertise resources. The partnership is providing needed analysis and improving productivity within the company while minimizing costs associated with software and expertise acquisition and retention. Future improvements to the DSS will likely include the addition of outbound supply chains (South American, Europe,) to the model and and use of the web-based capabilities of Simul8's Player.

Flexibility Mapping : Practitioner's Perspective

Although discrete-event simulations (DES) have been used extensively in business for decades to test the outcome of specific types of events, they have not traditionally been used in supply chains to optimize material flow. This paper demonstrates how an academician-company partnership was used to develop and support a DES decision support system in this context. Specifically, the academic partner developed and maintained a simulation model for optimizing load shipment wait-time and size into the Asian-Pacific region using data provided by a company. Both the simulation modeling process and the academician-industry partnership allow firms to create and maintain flexibility while reducing required in-house resources. For example, partnered firms need only ship new datasets to their academician partner when changes arise without permanently acquiring DBS-dedicated talent,

Reflecting Applicability in Real Life

This paper uses a real life situation to illustrate the way in which a discrete-event simulation was used by an academician-firm partnership to provide a flexible model designed to determine optimal shipment load/size/ time. The process described in the paper is viable across numerous situations, not just shipment size/time issues, and across most industries.

References

Albino V. and Garavelli A.C. (1999) Limited Flexibility in Cellular Manufacturing Systems: A Simulation Study. International Journal of Production Economics, 60/61, 447-455.

An C., Buckley S., Connors D., Feigin G. and Jayaraman R. (1994) Modeling and Simulation of Manufacturing and Distribution Networks, Proceedings of the TIMS/ORSA Conference, Boston.

Borenstein D. (2000) Implementation of An Object-Oriented Tool for the Simulation of Manufacturing Systems, International Journal of Production Research, 38 (9), 2125-2142.

Caprihan R. and Wadhwa S. (1997) Impact of Routing Flexibility on the Performance of An FMS A Simulation Study, International Journal of Flexible Manufacturing Systems, 9 (3), 273-298.

Cho H. and Wysk R.A. (1993) A Robust Adaptive Scheduler for an Intelligence Workstation Controller, International Journal of Production Research, 31 (4), 771-789.

Drake G.R., Smith J.S. and Peters B.A. (1995) Simulation as a Planning and Scheduling Tool for Flexible Manufacturing Systems, Proceedings of the Winter Simulation Conference, Arlington, VA.

Erickson C., Vandenberge A. and Miles T. (1987) Simulation, Animation, and Shop-Floor Control, Proceedings of the Winter Simulation Conference, Atlanta.

Ferguson L.A. (1998) Improving Global Supply Chain Management Decisions, Proceedings of the INFORMS Conference, Seattle.

Garg S., Vrat P. and Kanda A. (2001) Equipment Flexibility Vs. Inventory: A Simulation Study of Manufacturing Systems, International Journal of Production Economics, 70 (2), 125-143.

Gupta Y.P. and Goyal S. (1992) Flexibility Trade-offs in a Random Flexible Manufacturing System: A Simulation Study, International Journal of Production Research, 30 (3), 527-557.

Harmonosky C.M. and Barrick D.C. (1988) Simulation in a CIM Environment: Structure for Analysis and Real-Time Control. Proceedings of the Winter Simulation Conference, San Diego.

Harmonosky C.M. (1990) Implementation Issues Using Simulation for Real-Time Scheduling, Control, and Monitoring, Proceedings of the Winter Simulation Conference, New Orleans.

Jones A., Rabelo L. and Robohn S.F. (1995) A Hybrid Approach for Real-Time Sequencing and Scheduling, International Journal of Computer Integrated Manufacturing, 8 (2), 145-154.

Law A. and Kelton W. (2000) Simulation Modeling and Analysis, McGraw-Hill, New York.

Nandkeolyar U. and Christy D.P. (1992) An Investigation of the Effect of Machine Flexibility and Number of Part Families, International Journal of Production Research, 30 (3), 513-526.

Pflughoeft K. and Hutchinson O.K. (1999) The Effects of Workstation Flexibility on the (NYN) Job Shop Problem, International Journal of Flexible Manufacturing Systems, 11 (1), 103-113.

Pritsker A. (1995) Simulation and SLAM II, John Wiley, New York.

Shapiro J. F. (2001) Modeling the Supply Chain, Duxbury, United States.

Smith J.S., Wysk R.A., Sturrock D.T., Ramaswamy S.E., Smith G.D. and Joshi S.B. (1994) Discrete Event Simulation for Shop Floor Control, Proceedings of the Winter Simulation Conference, Orlando.

Thompson M.B. (1994) Expanding Simulation beyond Planning and Design, Industrial Engineering, 26 (10), 65-67.

Winston W. (2001) Simulation Modeling Using @RISK, Duxbury, United States.

Wu D.S. and Wysk R.A. (1989) An Application of Discrete-Event Simulation to On-Line Control and Scheduling in Flexible Manufacturing, International Journal of Production Research, 27, 1603-1624.

Yen-Hung-C. (1999) Global Supply Chain Simulation, Proceedings of the INFORMS Conference, Cincinnati.

Andrew A. Tiger, Ph D

Assistant Professor

Department of Management and Marketing

1405 N. 4th Ave. PMB 4152

Durant, OK 74701-0609

[email protected]

Penny Simpson, D. B A

Assistant Professor

Department of Management and Marketing

College of Business Administration

University of Texas Pan American

Edinburg, Texas 78539

Andrew Tiger is an Assistant Professor in the School of Business at Southeastern Oklahoma State University in Durant, Oklahoma. Prior. Andrew has worked for Minnesota Mining and Manufacturing, E. & J. Gallo Winery, Aspen Technology, and Texas A&M University - Kingsville.

Penny Simpson is an Assistant Professor of Marketing at the University of Texas Pan American. Results of her supply chain research have appeared in a number of journals including the Journal of Business Research, the Journal of the Academy of Marketing Science, the Journal of Marketing, and the Journal of Marketing Channels. She has also authored a chapter on target marketing and segmentation in a marketing principles textbook, Marketing: Best Practices.

Copyright Global Institute of Flexible Systems Management (GIFT) Oct-Dec 2003
Provided by ProQuest Information and Learning Company. All rights Reserved