Op t i m i za t i o n i n SA P

Su p p l y Ch a i n M a n a g e m e n t

Heinrich Braun,
Thomas Kasper
SAP AG

Agenda

Optimization in SAP SCM

Acceptance of Optimization

Summary

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

I n t r o d u c t i o n : Su p p l y Ch a i n M a n a g e m e n t

Supply Chain Management: Set of approaches utilized

to integrate suppliers, manufactures, warehouses and stores

so that merchandise is produced and distributed at right

quantity
locations
time

in order to minimize cost while satisfying service level requirements

[Simchi-Levi et al 2000]

Prerequisite: Integrated Supply Chain Model

Customers
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

DCs

Plants

Supplier

Su p p l y Ch a i n M a n a g e m e n t m y SA P SCM
Measure

Partner

Plan

Supply Chain Design

Source

Demand and
Supply Planning

Make

Direct
Procurement

Manufacturing

Deliver
Order
Fulfillment

Customer

Collaborate

Network

Collaborate

Private
Trading
Exchange

Supply Chain Collaboration

Supplier

Strategize

Supply Chain Collaboration

Supply Chain Performance Management

Private
Trading
Exchange

Network

Partner

Track

Supply Chain Event Management

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

m y SA P SCM : Pl a n n i n g L e v e l s
Strategic

Strategize
Supply Chain Design

Tactical

Plan
Demand and
Supply Planning

Operational

Source

Make

Direct
Procurement

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

Manufacturing

Deliver
Order
Fulfillment

Ch a l l e n g e : M a x i m i ze t h e a p p l i c a t i o n o f Op t i m i za t i o n i n SCM

Dos

Model Planning Problems as optimization problems

Use the best optimization algorithms

Respect the given run time for solution

Dont dos

Restrict the modeling to optimal solvable instances

Size

Constraints

Restrict your optimization algorithms to exact approaches neglecting

Iterated Local Search

Evolutionary Algorithms etc.

Our Goal
Leverage optimization algorithms by
Aggregation

Decomposition
for solving the Planning and Scheduling problems in SCM
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

Op t i m i za t i o n - Ex p e c t a t i o n b y o u r c u s t o m e r

Optimal Solution ?

At most 5% above Optimum ?

Best-of-Breed = Efficiency Acceptance !

Depending on problem complexity (model, size)

Given run time

Efficiency

Scalability

Decomposition = quasi linear

Toolbox: alternative algorithms
Parallelization

Return On Investment (ROI)

SAP AG 2002,

CP AI OR 2004,

Solution Quality
Total Cost of Ownership (Licenses, Maintenance, Administration, Handling)

Heinrich Braun

Pr o b l e m v e r s u s A l g o r i t h m s
OR-Researcher: accepts Problems, which solves his algorithm optimally
OR-developer: accepts Problems, which solves his algorithm efficiently
OR-user: accepts algorithms, which solves his problem effectively

ModelSpecification

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

OptimizationAlgorithms

Op t i m i ze r - Ex p e c t a t i o n b y o u r c u s t o m e r
Idealist

Searching for the global Optimum

Realist

Improving the first feasible solution until Time Out

Sisyphus

Knows, that the problem has changed during run time

Pragmatic

Solves unsolvable problems

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

Ch a l l e n g e : M a s t e r i n g t h e p l a n n i n g c o m p l e x i t y o f SCM

Hierarchical planning
Global optimization using aggregation (Supply Network Planning)

Feasible plans by local optimization (Detailed Scheduling)

Integration by rolling planning schema

SAP AG 2002,

CP AI OR 2004,

Modeling is crucial !!

Heinrich Braun

10

H i e r a r c h i c a l Pl a n n i n g
Supply Network Planning (SNP)
Mid term horizon

Time in buckets (day, week,..)

Short term horizon

Time in seconds

Global optimization

Local optimization

Maximize profit

Disaggregate global plan

Decide

Where to produce

How much to produce

How much to deliver

How much capacities

Linear optimization (MILP)

SAP AG 2002,

Detailed Scheduling (DS)

CP AI OR 2004,

Heinrich Braun

11

Time: When to produce

Resource:
On which alternative resource

Optimize production sequence

Scheduling algorithms (GA, CP)

SN P Op t i m i ze r : M o d e l Ov e r v i e w
Produce

Discrete Lots
Minimal Lots
Fixed Resource Consumption
Piecewise linear Costs

Transport

Transport
Capacity

Discrete Lots
Minimal Lots

Piecewise linear Costs

Production
Capacity

PPM
Storage
Capacity

Products

Safety Stock

Handling-In
Capacity

Store

Handling-Out
Capacity

with Shelf life

Satisfy Demand

Procure

Piecewise linear Costs

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

12

Priority Classes
Delay Costs
Non-Delivery Costs

PP/DS Op t i m i ze r : M o d e l Ov e r v i e w
Time Windows

Distances

earliest starting time

due date, deadline
delay costs

(with calendars)
minimal
maximal

Resource Selection
Alternative resources
Resource costs

Setup
sequence dependent
setup costs

Unary
Resources

Product Flow

Storage
resources

discrete
continuous

Multi-Cap
Resources

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

13

Ex a m p l e : Co m p l e x M o d e l i n g

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

14

I n t e g r a t i o n b e t w e e n SN P a n d PP/DS
SNP

Planning only in SNP horizon

Release SNP Orders only in PP/DS

horizon
Respect PP/DS orders as fixed
capacity reduction
material flow

PP/DS

Respect pegged SNP Orders as due

dates
No capacity reduction

But material flow

No restrictions for scheduling

PP/DS orders

SNP

PP/DS

PP/DS Horizon
SNP Horizon
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

15

I n t e g r a t i o n b e t w e e n SN P a n d PP/DS
SNP

Planning only in SNP horizon

Release SNP Orders only in PP/DS

horizon
Respect PP/DS orders as fixed
capacity reduction
material flow

PP/DS

Respect pegged SNP Orders as due

dates
No capacity reduction

But material flow

No restrictions for scheduling

PP/DS orders

SNP

PP/DS

PP/DS Horizon
SNP Horizon
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

16

Mast ering t he algorit hm ic c om plex it y: Aggregat ion

Model Accuracy versus Solution Quality -> SNP

Time Aggregation (telescopic buckets)
Limit the discretization

Product Aggregation (families of finished products)

Location Aggregation (transport zones, distribution centers)

However: Sufficient model accuracy of SNP for DS

Modeling of setup (important for process industry)

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

17

SN P Op t i m i za t i o n : A n t i c i p a t i o n o f Se t u p

Multi-Level Capacitated Lot Sizing Problem (MLCLSP)

Setup cost and/or consumption in each bucket
Good results
Setup consumption small compared to bucket capacity (small lots)

Bad results
Setup consumption big compared to bucket capacity (big lots)

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

18

SN P Op t i m i za t i o n : L o t Si zi n g i n SA P SCM
Proportional Lot Sizing Problem (PLSP)
Considering setup in preceding bucket

At most one setup per bucket

Constraints on cross-period lots (= campaign quantity)

Minimal campaign quantity

Campaign quantity integer multiple of batch size

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

19

SN P Ca m p a i g n Op t i m i za t i o n : B e n c h m a r k s

Comparison: real word problems of process industry

ca. 500 000 decision variables, 200 000 constraints
100,000000%
0

10

20

30

40

50

60

70

80

90

100

10,000000%

Optimality Deviation

1,000000%

0,100000%
PI-Bench2

0,010000%

PI-Bench7
PI-Bench8

0,001000%

PI-Bench9

0,000100%

0,000010%
0,000001%
0,000000%

time in minutes

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

20

M a s t e r i n g t h e a l g o r i t h m i c c o m p l e x i t y : De c o m p o s i t i o n

Global versus local optimality -> SNP + DS

Local optimality depends on neighborhood
High solution quality by local optimization
Local Optimization = Decomposition

Decomposition strategies

SNP: time, resource, product, procurement

DS: time, resource

Parallelization by Agents

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

21

T i m e De c o m p o s i t i o n - L o c a l I m p r o v e m e n t

Resources

Current window

Gliding window script

1. Optimize only in current window
2. Move window by a time delta
3. Go to first step
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

22

Time

SN P Pr o d u c t De c o m p o s i t i o n

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

23

SN P Pr o d u c t De c o m p o s i t i o n

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

24

SN P Pr o d u c t De c o m p o s i t i o n

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

25

SN P Pr o d u c t De c o m p o s i t i o n
100,00
90,00

Delivery [%]

80,00
70,00
60,00
50,00
40,00
30,00
LP Relaxation
Decomposed MIP
Global MIP

0,25

99,17

99,17

99,17

99,17

94,40

98,98

99,07

99,09

45,1162

53,33

62,81

65,13

Solution time [hrs]

Total cost

200.000.000,00

150.000.000,00

100.000.000,00

OPAL Executable
CPLEX 8.0

50.000.000,00

0,00

0,25

LP Relaxation

34.075.500,00

34.075.500,00

34.075.500,00

34.075.500,00

Decomposed MIP

63.231.900,00

38.170.900,00

37.871.200,00

37.860.100,00

2,44E+08

215.404.000,00

195.649.000,00

187.900.000,00

Global MIP

Solution time [hrs]

CP AI OR 2004,

Production Grouping
~15 comparable selections
179086 Variables
66612 Constraints
10218 Binary variables

250.000.000,00

SAP AG 2002,

Customer model

Heinrich Braun

26

Pentium IV 1,5 GHz

1 GB Main memory

M u l t i A g e n t Op t i m i za t i o n
Objective

Multi Criteria Optimization

250

200

user selects out of solutions with

Delay
Setup
Quality= D+S

similar overall quality

different components

Use power of Pallelization

150

Multi Agent Strategy

Different AGENTS focusing on Setup or Delay

or Makespan

100

Agents may use any optimizer (CP, GA, ..)

New solutions by local improvement

Integrated in Optimizer Architecture

50
0

SAP AG 2002,

Solut. Solut. Solut. Solut.

1
2
3
4
CP AI OR 2004,

Heinrich Braun

27

Performance

Speedup | available processors

Setup

Se v e r a l So l u t i o n s

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

Delay

28

Sc h e d u l i n g Op t i m i ze r A r c h i t e c t u r e
LiveCache

GUI

Time

Model Generator

Resource

Reporting

Core Model

Checking

Multi Agent

Control

Decomposition

Constraint
Programming

Genetic
Algorithm

Sequence
Optimizer

Basic Optimizer

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

29

Campaign
Optimizer

Op t i m i zi n g w i t h T i m e De c o m p o s i t i o n
LiveCache

GUI

Time

Model Generator

Product

Reporting

Core-Model

Checking

Control

Decomposition

SNP

Deployment

Basic-Optimizers

SAP AG 2002,

CP AI OR 2004,

Resource

Heinrich Braun

30

Vehicle
Allocation

Op t i m i zi n g w i t h T i m e De c o m p o s i t i o n

solved

solved

merge

TimeDecomposition

3-6
extract
store

solve

SNP

LP/MILP

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

31

Op t i m i zi n g w i t h T i m e De c o m p o s i t i o n

solved

solved

solved

solved

merge

TimeDecomposition

5-6
extract
store

solve

SNP

LP/MILP

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

32

Op t i m i zi n g w i t h T i m e De c o m p o s i t i o n

solved

solved

solved

solved

solved

solved

merge

TimeDecomposition

6
extract
store

solve

SNP

LP/MILP

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

33

Op t i m i za t i o n Pe r f o r m a n c e

Supply Network Planning

Pure LP

Without discrete constraints

Up to several million decision
variables and about a million
constraints
Global optimum guaranteed

For discrete constraints

No global optimum guaranteed

Quality depends on run time and
approximation by pure LP

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

34

Detailed Scheduling

Up to 200 000 activities

(no hard limitation)

First solution

as fast as greedy heuristics

More run time

improves solution quality

Su m m a r y - M a s t e r i n g t h e A l g o r i t h m i c c o m p l e x i t y
Aggregation

SNP: time, product, location

Automated Generation: SNP model deduced from DS model !!

SNP: time, resource, product, procurement

DS: time, resource

Parallelization by Agents

SNP DS

Decomposition

Hierarchy of Goals (customized by scripts)

SNP: Service Level Production Costs

DS: Service Level Storage Costs

Open to partner solutions: Optimizer extension workbench

Our Mission


Optimization as an integrated standard software solution

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

35

Fo c u s o f De v e l o p m e n t
In the past

larger optimization problems

better solution quality

more functionality

Current customer view

Very confident with performance, solution quality, functionality

Problem: Mastering the solution complexity (e.g. data consistency)

Current focus of development

More sophisticated diagnostic tools

Parallelization by GRID Computing

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

36

Agenda

Optimization in SAP SCM

Acceptance of Optimization

Summary

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

37

A c c e p t a n c e Cr i t e r i a

Heuristic (relaxing capacity) versus Optimization

Or

Why do not all SNP-Customers use optimization?

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

38

Model Ac c urac y

SNP Heuristic

Interactive Planning: Planner responsible for feasibility

limited model accuracy sufficient

SNP Optimizer

Automated Planning: Optimizer responsible for feasibility

high model accuracy necessary

Application in complex scenarios

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

39

A c c e p t a n c e c r i t e r i a : So l u t i o n q u a l i t y
SNP Optimierer

SNP Heuristik
LP:

Scaling of

&

Model size
Scaling of

Limited model accuracy

MILP: using decomposition

Very detailed
MILP: using decomposition

Model accuracy
Acceptance by
end user

&

simple model +
Transparency of
algorithms
acceptable solution

Not sufficient:
(Global) integer gap as Indicator
Not acceptable:
apparent improvement potential

Producing too early

Neglecting Priorities

Goal: Local Optimality

Decomposition

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

40

A c c e p t a n c e c r i t e r i a : So l u t i o n t r a n s p a r e n c y
SNP Heuristic

Solution stability

Sequential processing of
orders and few constraints
robust for small data
modifications

SNP Optimizer
Global Optimization
small data modifications may
cause large solution changes

MILP: change of search

No Black-Box
algorithms transparent

Interaction
Data/Model
versus
Solution

SAP AG 2002,

CP AI OR 2004,

Unforeseeable effects:
by considering both

globally and simultaneously

costs and constraints

Soft versus Hard Constraints

Heinrich Braun

41

violation depending on
solution quality!

Question: violation avoidable

by more run time?

Ac c ept anc e by different roles

OR Specialists

solution quality
run time behavior

IT department (TCO = Total Cost of Owner Ship, ROI = Return on Investment)

Maintenance
Administration
Integration
Upgrades / Enhancements

Consultants

Training / Documentation / Best practices

Complete and validated scenarios
Sizing tools

End user (Planner)

Modeling capabilities
Solution quality
acceptable solution in given run time, error tolerance
Solution transparency
Diagnostics, Warnings, Tool tips in case of errors

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

42

Op t i m i za t i o n a s s t a n d a r d s o f t w a r e ?!
Tradeoff

Solution quality
Solution costs (development, maintenance, ..)

Problems for special solutions

Modeling as moving target
Changing

master data

z additional constraints

z additional resources (machines)

Changing

transactional data

z Number of orders

z Distribution of orders to product groups

Changing

objectives (depending on the economy)

Effort for optimization algorithms only <10% of overall costs

Integration
Interactive

Planning
Graphical user interface
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

43

Op t i m i za t i o n a l g o r i t h m s
Which algorithms for which planning levels?

Supply Network Planning

LP / MILP Solver (CPLEX)

Classical Operations Research

Detailed Scheduling

Constraint Programming
Evolutionary Algorithms
Metaheuristics

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

44

Di a l e c t i c o f n a m i n g d e b a t e

Metaheuristics

Optimization

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

Heuristic

45

Op t i m i za t i o n n a m i n g c o n v e n t i o n

Academic: Optimization as an algorithm

find an optimal solution

or

Practical: Optimization as a dynamic process

the path is the goal

objective
function

time out

time
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

46

Op t i m i za t i o n i n p r a c t i c e
Global optimal

Often missed

Tradeoff: Model accuracy <-> solution quality

Metaheuristics

Very robust

Quality scales with given run time

Local optimal

A must have

The planer may not find simple improvements

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

47

Ac c ept anc e by different roles

OR Specialists

solution quality
run time behavior

IT department (TCO = Total Cost of Owner Ship, ROI = Return on Investment)

Maintenance
Administration
Integration
Upgrades / Enhancements

Consultants

Training / Documentation / Best practices

Complete and validated scenarios
Sizing tools

End user (Planner)

Modeling capabilities
Solution quality
acceptable solution in given run time, error tolerance
Solution transparency
Diagnostics, Warnings, Tool tips in case of errors

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

48

Op t i m i ze r a r c h i t e c t u r e
Acceptance-Criteria: ROI
Total Cost of Ownership
Licenses

(development effort)

Maintenance
Enhancements

Integration (to system landscape at customer)

Planning quality

Architecture
Toolbox

Exchangeable Components

Idea: Divide and Conquer for improving the components

SAP Netweaver

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

49

T o o l b o x : a l t e r n a t i v e Op t i m i ze r

Evolution / Competition of Optimizer

cf. Linear Optimization (ILOG CPLEX)

Primal
Dual

Simplex

Simplex

Interior

Point Methods

Alternative Optimizer for Scheduling

Constraint

Programming (ILOG)

Evolutionary
Relax

Algorithms (SAP)

and Resolve (University of Karlsruhe)

Alternative Optimizer for Vehicle Routing and Scheduling

Constraint

Programming, Tabu Search (ILOG)

Evolutionary

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

Algorithms (SAP)

50

Ac c ept anc e by different roles

OR Specialists

solution quality
run time behavior

IT department (TCO = Total Cost of Owner Ship, ROI = Return on Investment)

Maintenance
Administration
Integration
Upgrades / Enhancements

Consultants

Training / Documentation / Best practices

Complete and validated scenarios
Sizing tools

End user (Planner)

Modelling capabilities
Solution quality
acceptable solution in given run time, error tolerance
Solution transparency
Diagnostics, Warnings, Tool tips in case of errors

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

51

A c c e p t a n c e i n Co n s u l t i n g
Problem: Know How

Consultants sell methods to the customer which they can

master

modify

(additional heuristics)

Consequence for Optimization?!

Expert-Consulting Service for Optimization

Remote Consultant located in development

In particular, expert help for consulting partner

Quick Questionnaire

Early watch for planning complexity

Simple modeling tool
What

can I do to reduce model complexity?

Fallback Model: Which simplified model is uncritical?
Which model details are critical?

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

52

Si zi n g f o r Su p p l y N e t w o r k Pl a n n i n g
SN P Opt imizer Quest ionnair e
Continuous model

Resulting
variables

Estimate

Number of planning periods

Number of (planning relevant) location product combinations
Number of transportation lane product combinations
Number of location product combinations with customer demands
Number of location product combinations with corrected demand forecasts
Number of location product combinations with forecasted demands
Average lateness (in periods) allowed for demand items
Number of location product combinations with safety stock requirements
Number of location product combinations with product-specific storage bound
Number of location product combinations with an active shelf life
Number of production process models (PPMs)
Number of resources (production, transportation, handling in/out, storage)

82
14.500
176.000
11.705
0
11.705
1
11.705
0
1.641
50
Number of variables:
Number of constraints:

Discrete model

Discretized
periods

Number of PPMs with minimum lot size requirements

Number of PPMs with discrete lot size requirements
Number of PPMs with fixed resource consumption
Number of production resources with discrete expansion
Number of transportable products with minimum lot size requirements
Number of transportable products with discrete lot size requirements
Number of piecewise-linear cost functions (procurement, production, transport)

1.641
1.641
1.641

Additional number of discrete variables:

Additional number of constraints:

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

53

Resulting
constraints

1.189.000
14.432.000
1.919.620
0
1.919.620

1.189.000

959.810
0
0
134.562
4.100

959.810

959.810
0
959.810

0
4.100

20.558.712
4.072.530

Additional
discrete variables
134.562
134.562
134.562
0
0
0
0
403.686
0

Additional
constraints

Si zi n g f o r Su p p l y N e t w o r k Pl a n n i n g
SN P Opt imizer Quest ionnair e
Continuous model

Resulting
variables

Estimate

Number of planning periods

Number of (planning relevant) location product combinations
Number of transportation lane product combinations
Number of location product combinations with customer demands
Number of location product combinations with corrected demand forecasts
Number of location product combinations with forecasted demands
Average lateness (in periods) allowed for demand items
Number of location product combinations with safety stock requirements
Number of location product combinations with product-specific storage bound
Number of location product combinations with an active shelf life
Number of production process models (PPMs)
Number of resources (production, transportation, handling in/out, storage)

13
14.500
176.000
11.705
0
11.705
1
11.705
0
1.641
50
Number of variables:
Number of constraints:

Discrete model

Discretized
periods

Number of PPMs with minimum lot size requirements

Number of PPMs with discrete lot size requirements
Number of PPMs with fixed resource consumption
Number of production resources with discrete expansion
Number of transportable products with minimum lot size requirements
Number of transportable products with discrete lot size requirements
Number of piecewise-linear cost functions (procurement, production, transport)

1.641
1.641
1.641

Additional number of discrete variables:

Additional number of constraints:

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

54

188.500
2.288.000
304.330
0
304.330
152.165
0
0
21.333
650

Resulting
constraints
188.500
152.165
0
152.165
152.165
0
650

3.259.308
645.645

Additional
discrete variables
21.333
21.333
21.333
0
0
0
0
63.999
0

Additional
constraints

Si zi n g f o r De t a i l e d Sc h e d u l i n g
PP/ DS Opt imizat ion Q uest ionnaire
Version 4.0.1 for SAP APO 3.0, 3.1 and 4.0

Memory consumption

number of

Orders inside the optimization horizon

Average number of activities in an order
Average number of modes of an activity
Number of resources
Number of setup matrices
Maximum number of setup attributes in a setup matrix

5.000
10,0
3,0
10
2
500

objects
5.000
50.000
150.000
10
2
250.000

memory per

memory

object[kb] consumption[Mb]
4,50
22
5,00
244
1,00
146
530,00
5
55,00
0
0,10
49
8
130%
617

Environment [Mb]:
Safety factor:
Memory consumption [Mb]:

Complexity of the model

Size
Backward Scheduling
Block planning on resources
Bottleneck optimization
Campaigns
continous requirement/receipt
Deadlines modeled as hard constraints
Validity periods of orders
Cross-order relationships
Mode linkage
Sequence dependent setups activities
Shelf life or max. pegging arcs
Container resource
Synchronization on resources
Time buffer

yes
no
no
no
no
no
no
no
no
yes
no
no
no
no

degree of
difficulty
A
C
A
C
B
B
B
B
B
C
C
A
A
C
C

no of features
per degree of difficulty
A: 1
B: 0
C: 2
weighted sum 140

7.000
Degree
of complexity:
high

If the degree of complexity is not "moderate" SAP recommends the optimization consulting service

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

55

Si zi n g f o r De t a i l e d Sc h e d u l i n g
PP/ DS Opt imizat ion Q uest ionnaire
Version 4.0.1 for SAP APO 3.0, 3.1 and 4.0

Memory consumption
Orders inside the optimization horizon
Average number of activities in an order
Average number of modes of an activity
Number of resources
Number of setup matrices
Maximum number of setup attributes in a setup matrix

number of
25.000
10,0
3,0
10
2
500

objects
25.000
250.000
750.000
10
2
250.000

memory per

memory

object[kb] consumption[Mb]
4,50
110
5,00
1.220
1,00
732
530,00
5
55,00
0
0,10
49
8
130%
2.762

Environment [Mb]:
Safety factor:
Memory consumption [Mb]:

Complexity of the model

Size
Backward Scheduling
Block planning on resources
Bottleneck optimization
Campaigns
continous requirement/receipt
Deadlines modeled as hard constraints
Validity periods of orders
Cross-order relationships
Mode linkage
Sequence dependent setups activities
Shelf life or max. pegging arcs
Container resource
Synchronization on resources
Time buffer

yes
no
no
no
no
no
no
no
no
yes
no
no
no
no

degree of
difficulty
A
C
A
C
B
B
B
B
B
C
C
A
A
C
C

no of features
per degree of difficulty
A: 1
B: 0
C: 2
weighted sum 140

7.000
Degree
of complexity:
challenging

If the degree of complexity is not "moderate" SAP recommends the optimization consulting service

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

56

Ac c ept anc e by different roles

OR Specialists

solution quality
run time behavior

IT department (TCO = Total Cost of Owner Ship, ROI = Return on Investment)

Maintenance
Administration
Integration
Upgrades / Enhancements

Consultants

Training / Documentation / Best practices

Complete and validated scenarios
Sizing tools

End user (Planner)

Modeling capabilities
Solution quality
acceptable solution in given run time, error tolerance
Solution transparency
Diagnostics, Warnings, Tool tips in case of errors

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

57

Model Analysis
Product properties
Cardinality

of sourcing alternatives for a location-product

Maximum throughput per sourcing alternative for a location product
Lead-time per location product
Cycle analysis

Resource properties
Key

resources
number of (final) products requiring a resource

Cost model properties

Non-delivery

penalty configuration
non-delivery penalty > production costs+ Delay costs ?!
Transport-Storage
transport_costs > storage_costs ?!
Push analysis
storage_costs_L1 + transport_costs < storage costs_L2 ?!

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

58

Pu l l /Pu s h Re d u n d a n c y : Ex a m p l e Pu l l Cl o s u r e

Demand

Eliminate location-products not required for demand- or safety stock

satisfaction, min resource utilization, or not involved in cost push
Eliminate all dependent lane-products, lanes and locations

Eliminate all dependent activities (procurement, transport, PPM)

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

59

Co m p u t a t i o n a l T e s t s : Ph a r m a I n d u s t r y
Problem characteristics
#buckets #loc-products #lane-products #ppm
#products #locations #lanes
#ineq
#vars
memory consumption
24
8.788
2.647
10.680
5.644
64
207
201.327 524.578
994.050.048

Level of redundancy
Memory Consumption

Data Reduction
100%

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%

50%
0%
w ith cost
push

w ith cost push

t
t
uc
d
uc
o
d
r
p
ro
n-p
e
it o
a
an
#l
oc
l
#

pm
#p

ct
du
o
r
#p

w /o cost push
SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

w /o cost push

oc
#l

rc
#a

w ith cost push

60

1
w /o cost
push
w /o cost push

Co m p u t a t i o n a l T e s t s : Co n s u m e r Pr o d u c t s
Problem characteristics
#buckets #loc-products #lane-products #ppm
#products #locations #lanes
33
5.092
23.107
1.169
260
32
146

#ineq
#vars
201.327 524.578

memory consumption
1.569.828.864

Level of redundancy
Data Reduction

Memory Consumption

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%

100%
80%
60%
40%
20%
0%

CP AI OR 2004,

Heinrich Braun

#a
rc

#l
oc

du
ct

#p
ro

#p
pm

#l
oc
at
io
npr
od
#l
an
uc
et
pr
od
uc
t

w/o cost push

w/o cost push

SAP AG 2002,

w ith cost push

with cost push

61

w /o cost push

w /o cost push

w ith cost push

Re p o r t o f So l u t i o n Qu a l i t y
Key Indicators

Resource utilization
Setup
Idle

Stock

time

time

Safety
Range

Demand

of coverage

Backlog
Delay
Non

(Percentage of volume, average of delay in days, priorities)

Deliveries

Aggregation Drill Down

Location

Product

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

62

Advanc ed solut ion analysis

Explanation

Non deliveries

Delay

Missed safety stock

Easier Optimizer Customizing

Preconfigured optimizer profiles

Support for generating cost parameters

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

63

Su m m a r y A c c e p t a n c e o f Op t i m i za t i o n i n Pr a c t i c e
Acceptance by OR-Specialist

Evolutionary algorithms are also optimization algorithms

Bridging the gap between Local Search and MILP

Acceptance by IT-department / Decider of the investment

Solution completeness, Integration, Openness to Enhancements

Component Architecture (SAP Netweaver)

Acceptance by Consultant

Know How, References, Best Practices, Quality guarantees

Optimization as commodity needs patience

Acceptance by End user

Explanation Tools

Semi-automatic optimizer configuration

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

64

Re a l l o c a t e Or d e r s u s i n g a g g r e g a t e d l e v e l (SN P )

Problem
Scheduling a late order or Solving a machine break down

Solving on same location and same production process model

Re-optimize in detailed scheduling

Change location or production model

Remove orders overlapping machine break down

Resolve problem on aggregate level (SNP)

Selecting appropriate location and/or alternative PPM

Release order to detailed scheduling

Re-optimize detailed scheduling

ONLY APPLICABLE: Inside overlapping horizon of SNP and PP/DS

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

65

SN P c o n s i d e r s s e t u p i n PP/DS
Setup for each bucket
Enhancing PP/DS orders by SNP orders

Bucket 1
SNPorders

PP/DS
orders

Bucket 2
B

SNP

PP/DS

SAP AG 2002,

CP AI OR 2004,

Heinrich Braun

66