Applications for stochastic models in lean management … · 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9...

KIT – Universität des Landes Baden-Württemberg und

nationales Großforschungszentrum in der Helmholtz-Gemeinschaft www.kit.edu

Institut für Fördertechnik und Logistiksysteme

Applications for stochastic models in lean management – Opportunities and missing links

GOR – AG Supply Chain Management

Synthese von Lean Management und Operations Research

Leinfelden, 27.9.2013 – Kai Furmans

© Institut für Fördertechnik und Logistiksysteme, 2009 2

IFL

Toyota Lean - Shock Lean-Revival

1950 1990 - 1995 2000 - today

Development of Lean Systems

MRP ERP SCM as APS

1960 – 1970 1970 – 1990 1995 – today

Lean

SCM

© Institut für Fördertechnik und Logistiksysteme, 2009 © Institut für Fördertechnik und Logistiksysteme, 2012 © Institut für Fördertechnik und Logistiksysteme, 2010

Why Lean Management?

3

Empirical and Anecdotal Evidence, that Lean

Management enhances Productivity more than the

average productivity gains in the base population

Leadership and

Target Setting

Design Elements

which tackle

variability and

create a direct

relation between

Design, KPI and

KPR

Associate

involvement and

Empowerment,

large base for

problem solving

See i.e. Dehdari 2013

© Institut für Fördertechnik und Logistiksysteme, 2009

Some foundations – Supply Chain Physics

Leadership and

Target Setting

Design Elements

which tackle

variability and

create a direct

relation between

Design, KPI and

KPR

Associate

involvement and

Empowerment,

large base for

problem solving

Stochastic

models can

explain, why

it works


IFL

Supply Chain Physics – Little‘s Law

stock = throughput * throughput time

Ns [units] = λ [units per time] * ts [time]

• Initial situation

• Double throughput with same throughput time

doubled stock

• Same throughput with longer throughput time

increased stock


Supply Chain Physics – Little‘s Law

Application in Value Stream Mapping

Response Time = Inventory ∗ Customer Takt

CT =

160 min = 1/3 d

Source: valuestreamguru


IFL

Supply Chain Physics –

Variability Generates Waiting Times

random

interarrival time

& clocked

processing

Waiting times

arise – visible by

stocks (Little)

arrivals &

processing

clocked

no waiting time


Ereignis

0,8

1

0 200 400 600 800 1000 1200 1400 1600 1800

Ereignis

0,8

1

0 500 1000 1500 2000 2500

Excursus - Variability

2

2

2

2

)(

)(

XXE

XVarc

c²=0,01

c²=1,00

9

IFL


Supply Chain Physics –

Capacity Utilization (workload) and Variability

0

1

2

3

4

5

6

7

8

9

10

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

stock

workload

0

5

10

15

20

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

stock

workload

With the workload, the influence of

variability on stocks and leadtimes

increases

Comparison of the average stock of the

system at varying variabilities

Low variability Medium variability

High variability

No variability

Medium vs. high variability at 94% workload

Low vs. medium variability at 94% workload

No vs. Low variability at 94% workload

10

IFL

21

222

ba ccN


IFL

Typical sources for Emergence and

Intensification of Volatility

Caused by the design of the material flow

Caused by downtimes

In the material flow

In the processing

As a result of scrap

Caused by forming batches in containers

Caused by inventory control


IFL

Branches in the Production Flow Generate

Variability Initial

variability

in front of

the branch

Whitt, 1983

Create Linear flow


Oversized Container Fill-Up Capacities

Generate Variability

0,00

50,00

100,00

150,00

200,00

250,00

1 5 9 13 17 21 25 29 33 37 41 45 49

dem

an

d / o

rder

qu

an

tity

period

Demand

Order

Average demand: 100 / period, = 20

13

IFL

0

20

40

60

80

100

120

0 50 100 150 200 250

sta

nd

ard

de

via

tio

n d

em

an

d

container quantity

standard deviation container calls (demand)

Standardabweichung Behälterabrufe

Standard

deviation

container calls


IFL

00,05

0,1

0,15

0,2

0,25

0,3

0,35

0,4

0,45

0,5

0,55

0,6

0,65

0,7

0,75

0,8

0,85

0,9

0,95

0

5

10

15

20

25

30

35

40

Auslastung

Ve

rw

eilze

it d

er A

ufträg

e

Effect of Downtimes

V ( t) =

Z t

0V ( u) du =

A( t)X

i = 1

³W i ¢B i + B 2

i =2´

+ ²( t )

( 1)

The increase in

leadtime caused by

downtimes is

essentially higher

than it can be

explained only by the

loss of capacity!

With

ou

t d

ow

ntim

es

dw

ell

tim

e

Incre

ase

in

ca

pa

city

utiliz

atio

n (

wo

rklo

ad)

by d

ow

ntim

es

capacity utilization

(workload)


IFL

Methods of Control Effects

Why pull?

Why leveling?


IFL

Open vs. Closed Systems

1

N

Open system:

• control of system load (z.B. MRP)

• result is stocks

Closed system:

• control of stocks

• result is throughput (performance)

The required stock (leadtime at equivalent throughput) is in closed systems

always less than/equal to the one in opened


IFL

Kanban-Systems are Closed loops with a Brake

Pre-

Products

Demand

Pre-Products

Pre- and finished products

with Kanban (WIP)

Demand

satisfied

demand

Station of

Synchronization j


0

5

10

15

20

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

stock

workload

Impact of the Design Elements of Lean

Systems

18

Variability Poka Yoke

Standards

Small handling

units

Milkruns

SMED

Linear flow

Levelling

Load Control

Kanban

TPM

Low Cost Automation


Leadership and Target Setting

Leadership and

Target Setting

Design Elements

which tackle

variability and

create a direct

relation between

Design, KPI and

KPR

Associate

involvement and

Empowerment,

large base for

problem solving


Leadership and Target Setting

Evolution of Management Style

21 27.09.2013

Directions

Delegation

and motivation

Pressure by

results publication

Develop

target states

Creativity

z.B. size of company

Maturity

Age of organization Inspired by Greiner, HBR, 1972

© Institut für Fördertechnik und Logistiksysteme, 2009 © Institut für Fördertechnik und Logistiksysteme, 2012 8

IFL

Approach:

Design a System, which achieves your goals!

Target Value Stream Design

Business Targets in

€

translate from

physical units

to €

translate from

€ to physical units

Implemented Value Stream

Calculate

KPI and

KPR

KPI and KPR

met?

Improvement

Set New Targets

Implement


Support needed by OR – some examples

Develop exact or approximate methods, which allow a fast

and simple sizing and evaluation of value streams

Sizing supermarkets (here, only number of Kanbans)

23 27.09.2013

Source: vision-lean.com

Shingo:

𝐾 =𝑄+𝛼

𝑛

K = number of kanbans;

Q = quantity of products in batch production;

α = minimum security stock level;

n = quantity of products transported on a pallet.

Monden:

K = 𝑑 𝑡𝑒 + 𝑡𝑓

𝑐(1 + 𝛽)

k = number of kanbans;

d = demand on the planned period;

te = waiting time, defined from the time

since the necessity of production is defined

until effective production starting time;

tf = time it takes to produce a container

(one kanban) of products;

β = safety factor (around 15%);

c = container capacity.

Bosch:

K = 𝑅𝐸 + 𝐿𝑂 + 𝑊𝐼 + 𝐿𝑂 + 𝑆𝐴

Source: Salgado, Varela, 2010 …

DIMASCOLO, M. ; FREIN, Y. ; DALLERY, Y. :

An Analytical Method for Performance

Evaluation of Kanban Controlled Production

Systems, 1993


Dynamic Aspects

When to change the supermarket size?

Determine escalation levels -> when to change operations

When to increase / decrease capacity?

When to use a second source…

How to ramp up production over several levels?

Fill the buffers first?

Try to ramp up simultaneously?

What is the impact of adjustable capacity, when capacity

adjustment is subject to:

Early announcement?

Limitations in extend?

How much does (possibly inaccurate) advance information

help in a dynamic environment?

24 27.09.2013


IFL

Approach:

Design a System, which achieves your goals!

Design

Description and Calculation

Layout

Try it!

Measure it!


Continuous Improvement at the Base

Leadership and

Target Setting

Design Elements

which tackle

variability and

create a direct

relation between

Design, KPI and

KPR

Associate

involvement and

Empowerment,

large base for

problem solving


IFL

The Continous Improvement Cycle

Describe

Standards

Observe

Standards -

Measure

Deviations

Pareto

Analysis of

Deviations

Solve most

Severe

Problems

first

Improve

Standards

Location,

Time,

Sequence

Right Location

?

Right Time?

Right

Sequence?


IFL

Development of a Lean Operation

Shop-floor

level

Project

level

Management

level

Mind-set

Lean basics

training 5S, pull,

Kanban…

Value Stream

Analysis

Face-the-Facts On the shop floor

Awareness: we are

not good enough

t

Basics

training Kaizen, KVP, CIP

Problem Solving…

Value Stream

Design Vision, First Step,

Set-up Project

Face-the-Facts On the shop floor

Willingness: we have to change

Implementation

of basics Attention to

Work Systems

i.e. MTM…

Implementation upstream, from

customer to supplier

one value stream,

no shortcuts

Face-the-Facts

Attention to

Change Project

Support: we support those

who try to improve

Continuous

Improvement

Advanced Lean

i.e. heijunka

Hand Over responsibility moves

from project to

operations

Request

Adherence to

Defined

Processes

Continuous

Improvement: we are never satisfied

Warming Storming Norming Performing


Problem Solving Approaches

What is better:

Put more emphasis on problem solving or

just have more inventory?

What is better:

A simple system, which is well understood and under control or

an optimal system, which is not quite understood and not under

control?

What is the impact of reward systems on

Problem solving vs.

having excess capacity / excess inventory?

30 27.09.2013

© Institut für Fördertechnik und Logistiksysteme, 2009 31 27.09.2013

Karlsruher Institut für Technologie (KIT)

Institut für Fördertechnik und Logistiksysteme (IFL)

Postanschrift Kaiserstr. 12

76128 Karlsruhe

Besucheranschrift Gotthard-Franz-Str. 8 Geb. 50.38

76131 Karlsruhe

Telefon +49 (721) 608-48600

Telefax +49 (721) 608-48609

eMail [email protected]

www http://www.ifl.kit.edu

Berlin

Hamburg

Rügen

Fehmarn

Bremerhaven

Kiel

Schwerin

Rostock

Bremen

Flensburg

Lübeck

Hannover Wolfsburg

Braunschweig

Oldenburg

Wilhelmshaven

Osnabrück

Dessau

Salzgitter

Göttingen

Hamm

Münster BielefeldHildesheim

Halle

Dresden

Leipzig

ChemnitzZwickau

GeraJenaErfurt

Potsdam

Cottbus

Siegen

Mönchen-gladbach

Krefeld

Köln

HagenDüsseldorf

Bonn

Koblenz

Trier

Saarbrücken

Kaiserslautern

Darmstadt

Frankfurtam Main

WiesbadenMainz

Offenbach

Würzburg

Mannheim

HeilbronnKarlsruhe

Ludwigshafen Heidelberg

Pforzheim

Stuttgart

Freiburg

UlmAugsburg

München

Regensburg

ErlangenFürth

Nürnberg

PaderbornEssen

DortmundDuisburg

Ostfriesische ins.

Sylt

Konstanz

Applications for stochastic models in lean management … · 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9...

Documents

Transcript of Applications for stochastic models in lean management … · 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9...