Oliver Rose Dresden University of Technology Department of Computer Science Chair for Modeling and Simulation Benefits and Drawbacks of Simple Models for

Oliver Rose

Dresden University of TechnologyDepartment of Computer ScienceChair for Modeling and Simulation

Benefits and Drawbacks of Simple Models

for Complex Production Systems

2

Dresden University of Technology

Department of Computer Science

Oliver Rose

Simple Models


Dresden located in the state Saxony, often called “Silicon Saxony” (factories from Infineon, Qimonda, AMD, and ZMD; lots of suppliers)

Largest German University of Technology Focus on Computer Science, Electrical Engineering, and

Mechanical Engineering About 35,000 students, 3,000+ students in Computer

Science, 700+ beginners in winter semester 2005/2006 Faculty of Computer Science consists of 30 chairs Institute of Applied Computer Science has 5 chairs,

dealing with different aspects of factory planning, control, and automation

3



Oliver Rose

Simple Models

Research Overview

Complex Production System

Model

Simulation

Dispatching Simulation-based

Scheduling

Measu

rem

en

t &

An

aly

sis

Au

tom

ati

on

of

Data

Tra

nsfe

r

High-level Production Control

4



Oliver Rose

Simple Models

Cooperations

Industry Infineon Technologies AG, Munich and Dresden AMD Saxony LLC & Co. KG, Dresden KBA Planeta, Radebeul (Offset Printing Machines) Airbus Deutschland GmbH, Hamburg

Academia Arizona State University, Tempe, AZ, USA Singapore Institute of Manufacturing Technology Georgia Institute of Technolgy, Atlanta, GA, USA FernUni Hagen, Germany Fraunhofer IPA, Stuttgart, Germany

5



Oliver Rose

Simple Models

Introduction

Goal of the study: Estimate the transient behavior of

a wafer fab after recovery from a catastrophic

failure in order to improve the flow of material

Motivation: Infineon engineers reported that large

amounts of WIP are present in the fab even weeks after

end of repair

Goals of improvement:

reduction of WIP

reduction of cycle times

on-time delivery (reduction of variability of cycle

times)

6



Oliver Rose

Simple Models

Semiconductor Manufacturing

Back End

Front End

Wafer Fab

TestAssembly

Probe

7



Oliver Rose

Simple Models

Flow of material

Fotos: Fullman-Kinetics, Varian, Sematech International

8



Oliver Rose

Simple Models

Characteristics of wafer fabrication facilities

Large number of processing steps, typically several

hundreds

Large number of tools of different types: photo

equipment, ovens, etching equipment, ion implanters, …

Wafer are build up in layers: reentrant flow of material,

jobshop type way of production

Machine breakdowns (typical availability: 70-90%)

Auxiliary resources, e.g., reticles (photo masks)

Batch tools with complex batching criteria

Sequence dependent setups

9



Oliver Rose

Simple Models

Important performance measures

Cycle time: low

Output: high

Machine utilization: high

Inventory (work in process, WIP): low

Yield (percentage of good dies on a wafer): high

Cost per die: low

Conflicting goals!

10



Oliver Rose

Simple Models

Controllable factors

Factory load (wafer starts per week)

Product mix

Number of machines and operators

Preventive maintenance policies

Production planning & control policies:

scheduling vs. dispatching

lot release vs. shop-floor control

…

11



Oliver Rose

Simple Models

Fab Model

fixed number of cycles

bottleneckworkcenter

delay unit(rest of the fab)

finished?

yes

no

lot release

12



Oliver Rose

Simple Models

Modeling Details

Dispatching strategies:

FIFO

Critical Ratio

Slack Time

Delay time distributions of the delay unit

Lot release policy

Partial bottleneck breakdowns

13



Oliver Rose

Simple Models

Modeling the Rest of the Fab

?

14



Oliver Rose

Simple Models

Delay Time Distributions

Constant amount of time for sum of processing times

Distributions of different variability for sum of non-

processing times (waiting times, transport times, etc.):

Constant (coefficient of variation: 0.0)

Erlang-5 (coefficient of variation: 0.447)

Exponential (coefficient of variation: 1.0)

Independence of consecutive and parallel delays assumed

15



Oliver Rose

Simple Models

Constant Delay

FIFOCritical RatioSlack Time

time

WIP

16



Oliver Rose

Simple Models

Shifted Exponential Delay


time

WIP

17



Oliver Rose

Simple Models

Shifted Erlang Delay


time

WIP

18



Oliver Rose

Simple Models

Delay Distribution from Real Data

0

0.01

0.02

0.03

0.04

0.05

0.06

0 20 40 60 80 100 120 140

hours

datagamma

19



Oliver Rose

Simple Models

Partial Bottleneck Workcenter Breakdowns

?

20



Oliver Rose

Simple Models

Modeling of Partial Breakdowns

number ofbrokenmachines

4

2

1

duration of breakdowns

50 100 200

21



Oliver Rose

Simple Models

One Broken Machine for 200 Time Units

time

WIP


22



Oliver Rose

Simple Models

All machines broken for 50 time units


time

WIP

23



Oliver Rose

Simple Models

Combination of two dispatch strategies

?

Critical Ratio -> FIFOSlack Time -> FIFO

24



Oliver Rose

Simple Models

Combination of Critical Ratio and FIFO

FIFOCritical Ratiot<2500 CR, t>=2500 FIFO

time

WIP

25



Oliver Rose

Simple Models

Stop Lot Release During Breakdown Period

26



Oliver Rose

Simple Models

Lot Releases Stopped

FIFO (stopped)

Critical Ratio (stopped)

Slack Time (stopped)

FIFOCritical Ratio

Slack Time

time

WIP

27



Oliver Rose

Simple Models

Estimation of Cycle Time Distributions

0

0.005

0.01

0.015

0.02

0.025

0.03

0 100 200 300 400 500 600 700 800 900

hoursHours

0

0.005

0.01

0.015

0.02

0.025

0.03

0 100 200 300 400 500 600 700 800 900

hoursHours

0

0.002

0.004

0.006

0.008

0.01

0.012

700 750 800 850 900 950 1000 1050

hours

fullsimple

FullSimple

Hours

FIFO

0

0.01

0.02

0.03

700 800 900 1000

hours

fullsimple

FullSimple

Hours

CR

28



Oliver Rose

Simple Models

Conclusions

Simple models useful for understanding complex

production systems

Simple models useful for the design of simple and robust

scheduling heuristics

Not a tool for beginners

Pitfall of oversimplification

Simplification must not be the goal but only the method

to reach the goal

Keep the model as simple as possible but not simpler!

Documents

Oliver Rose Dresden University of Technology Department of Computer Science Chair for Modeling and Simulation Benefits and Drawbacks of Simple Models for