SAP APO Overview

SAP APO Overview
Page 1
MindTree Consulting Proprietary & Confidentia
SAP APO Architecture

A Solution to Provide An Integrated & Synchronized SC
Process
SAP Advanced Planner & Optimizer (APO)

DP
Business
Warehouse
(Reporting)
SNP
PP/DS
TP/VS
ATP
SC Cockpit
APO
Solvers
liveCache
APO
Application Link Enabling
Model Generator, Mapping, Connectivity

OLTP
R/3
Page 2
1997 SAP AG APO 2
R/3
Legacy
OLTP
Non-R/3
OLTP
Non-R/3
OLTP
MindTree Consulting Confidential
SAP APO Application Architecture
R/3
Supply Chain Cockpit
BIW
Advanced
Advanced Forecasting
Forecasting and
and Demand
Demand Planning
Planning
Historical
Data
Customer
order
Manufacturing
Execution
Inventory
Management
Page 3
ATP
ATP
Production
Planning and
Scheduling
Deployment
Deployment
Supply
Supply
Network
Network
Planning
Planning
Key
Perform.
Indicators
External
data
(e.g.POS)
Multiple System Environment

Legacy
Systems
Planning
Systems
Internet
&
EDI
ALE
WebGUIs
APO
infocubes
liveCache
SAPGUIs
R/3
4.x
BW
Page 4
R/3
3.x
liveCache
Production Process
Model
Optimized for realtime scheduling and

pegging (supply
demand, constraints)
Time Series
+10
+20
Optimized for
fast
response ATP
Page 5
-5
APO Solvers
liveCache
Model Generator,
Metaheuristics,
Optimizing Algorithms
Heuristics
Memoryresident data
object
processing
Supply Chain
Network
Application
Representation of
the extended supply
chain
liveCache
Application
Manages large amounts of data in main memory
Uses both relational and object-oriented functionality
(ADABAS)
Integration with R/3

Uses SQL Interface
Standard Transaction Handling

Locks, Rollbacks, Commits
Advantages (Performance)
Avoids disk I/O
References object via pointers
Stores complex data structures in object oriented data
structures
Recovers quickly in case of system crash
Page 6
liveCache
Page 7
A tool
A tool for processing large volumes
of data in main memory.
Main objective = higher performance
Avoid disk I/O
Stores optimized data structures
Scalability
number of processors within one liveCache
number of liveCaches
liveCache
Memoryresident data
object
processing
Application
liveCache
Main objective = higher performance:

Performance critical routines (in C++) are running in address space
of liveCache Management System
=> no heavy data transfer between application and data storage
Application
Application
Application
> 1 ms
Data
DataStorage
Storage
Page 8
< 10 s
Data Storage
liveCache
liveCache Integration with R/3
Forecast
DP
liveCache
Supply Plans
On Hand
In transits
Purchase Reqmts
Purchase Order
Sales Order
Sales Order
SD
Page 9
Production Order
Confirmation
Purchase Reqmts
Transport Reqmts Planned Order
Transport Orders
MM
PP
Optimizer
Optimizers are used for:
Supply Network Planning (SNP)
Production Planning / Detailed Scheduling (DPS)
Capable to Match (CTM)
Optimizer routines are developed in C++, which increases

the speed of the program
Only available on NT and Windows 2000
Communication between the Optimizer routines and APO
through SAP Gateway
Page 10
SAP APO Integration

For SAP R/3 instances, SAP provides the Core Interface Facility (CIF) which
dramatically simplifies integration to/from APO. Integration with non-R/3
systems is achieved through ALE and SAP provided BAPIs.:
APO
R/3
R/3
R/3
4.x
4.x
4.x
Supply Network
Planning
Supply
R/3
R/3
R/3
4.x
4.x
3.x
R/3
R/3
Non
4.x
4.x
R/3
Page 11
Demand Planning
A
T
P
Production Planning
/ Detail Scheduling
Chain
Cockpit
Live
Cache
APO - R/3 Integration
Core Interface Facility (CIF)

CIF is an online transaction that defines active data
channel(s) in R/3 for data transfer between R/3 Systems
and APO. It has the following features :
Real Time Interface
Determines Source and Target Systems within Complex System
Environments
Supplies APO with Relevant Master and Transaction Data
Forwards Data Changes (Transaction Data)
Returns Planning Results to SAP R/3
Page 12
CIF : Core Interface

APO Core Interface (CIF) is the communication layer to
be applied to R/3 to enable an exchange of data between
R/3 and APO.
APO-CIF is delivered as a plug-in . This is a general
product name given by SAP for the R/3 interfaces to the
new dimension applications.
Page 13
CIF Functions Architecture

ERP -> APO
Master Data
Transaction Data
Locations
Products
PPMs (BOM+Routing)
Characteristics
Capacities
Planned/Production
Orders
Sales Orders
Purchase Orders
Stocks
ATP Requests
APO -> ERP
Planning Results
ATP Results
Manufacturing Orders
Procurement Orders
VMI Sales Orders
APO
ERP
BW
APO
ERP
ERP
APO
Page 14
ERP
Technical Considerations
Page 15
Modules being Implemented

Multiple Parallel Rollouts
Distributed vs. Central Architecture
UNIX vs. NT
Single Client Strategy
Number / Location of Users
Amount of ALE Traffic
Volume of Data
SAP APO
Demand Planning
Page 16
Supply Chain Planning Cycle
Planning horizon
Demand Planning (DP)
Production Planning (PP)

Detailed Production Scheduling (DPS)
Deployment
Transportation Plan / Vehicle Sched.

Page 17
Planning Area Functionality

Defined like an InfoCube
Contains Characteristics and Key Figures
Maps where key figures for the planning area are
stored (InfoCube, Orders in liveCache, Time Series in
liveCache)
Planning Areas can be relevant for DP and SNP at
the same time
Forecast settings are done for planning area
Creation of time series objects
Page 19
Planning Area Flow

Planning objects
Key Figures
Sales
Production
Stock
Assign Key Figures

by Aggregate
Aggregate
Brand
Sales
Customer
Production
Details
APO Planning Version

Planning UOM
Time Bucket Profile
Page 20
Brand
Product
Customer
Sales
Production
Stock
Consistent Planning
Planning
at
io
re
g
Ag
g
ati
Lowest
Level
on
Page 21
Pro Rata
or
Proportional
Factors
g
re
gg
Proportion
al values
generated
sa
Di
Planning
Level
Statistical Toolbox
Univariate Forecasting
Moving Average
Simple Linear Regression
Exponential Smoothing
Holt-Winters
Crostons Model (for sporadic demand)
Causal Analysis
Multiple Linear Regression
Composite Forecasting
Weighted Averaging of Multiple Models
(Ex. Constant, Trend, Seasonal, MLR)
Page 22
Forecast Profiles
Profiles:
Assign a Planning Area
Define which key figure you
want to be forecasted
Define past and future
periods
Specify models to be used
for:
Univariate forecast
Multiple linear regression
Composite forecast
Page 23
Master
Forecast profile
Univariate profile
MLR profile
Composite profile
SAP APO Overview:

Supply Network Planning
Page 24
The Supply Chain: Original Supply

to Final Consumption
INFORMATION FLOW
Transfer
Transfer
Supplier
Manufacturing
Transfer
Distribution
Transfer
Retail Outlet
Consumer
CASH FLOW
Supply Chain Optimization
Increase Customer Responsiveness at Least Cost

Page 25
What is Supply Network

Planning ?
A planning approach to create Tactical Plans and
Sourcing Decisions that takes the complete supply
network into consideration
Supply Network
Planning
Deployment
Page 26
Transport
Load Builder
Meet Forecast and Actual Demand by:

Optimal use of Manufacturing, Distribution and
Transportation Resources
Consider all constraints in the supply chain
SNP Planning Functionality
Supply Network Planning Strategies
Heuristics
Optimization
Capable to Match
Propagation
SNP
PP/DS
Deployment
Fair share, push rules and deploy to order
Optimization
Deployment
Transport Load Builder

Leveling in transport loading
Page 27
TLB
APO Planning Functionality Sequence

Planning Horizon
Detailed Production Scheduling (DPS)
Deployment
Transport Load Builder (TLB)
Transportation Planning and Vehicle
Scheduling (TPVS)
Page 28
SNP Process Flow

Planning Area Administration
APO Master Data Setup
Model/Version Creation
Supply Chain Model Setup
Release of Demand Plan to SNP
Planning Method Determination
& Profile Settings
Transport Load Building
Simulation
Supply Network
Planning Run
Interactive Planning
Deployment Run
Management
by Exception
Conversion into
PP/DS orders
Release of SNP
Plan to DP
PP/DS Planning
Page 29
Examples : SNP Business uses

Beyond day-to-day planning
High level capacity planning and macro global scheming
to aid with capital equipment decision making based on
sales runs.
Support system for modeling what if scenarios that
would impact the bottom line.
Visibility for planning with confidence. System provides
cost- trend analysis for control on manufacturing
expansion, Out-sourcing, 3rd party supply, etc.
Page 30
Beyond Traditional DRP
Production Process
Model
Manufacturer
Plant
Manufacurer
DC
Customer DC
Supplier Plant
Supplier WH
Page 31
SNP Distribution Network
Transport Order
Planned Order / Procurement
Customer
Demands
Distribution
Centers
Plants
Suppliers
Page 32
Objects in the Supply Chain Network

Objects include:
Locations
Products
Resources
Production Process Models
Transportation Lanes provide links
between objects in a supply chain
model
Page 33
Vendor Managed Inventory

MANUFACTURER
RETAILER
APO
Plant
Inventory
Warehouse Inventory Distribution
(integrated)
(EDI 852)
Center
Forecast
(integrated)
SNP
R/3,
Sales order
Legacy
Page 34
Forecast
(EDI 830)
VMI
Order (EDI 855)

SNP Process Flow
Planning Area Administration

APO Master Data Setup
Model/Version Creation
Supply Chain Model Setup
Release of Demand Plan to SNP
Planning Method Determination
& Profile Settings
Transport Load
Building
Simulation
Supply Network
Planning Run
Interactive Planning
Deployment Run
Management
by Exception
Conversion into
PP/DS orders
Release of SNP
Plan to DP
PP/DS Planning
Page 35
SNP Planning Area Administration

- Process Flow
Page 36
Create Planning
Object structure
Characteristics
Create
Planning Area
Key figures
Initialise
Planning area
Attributes
Create
Planning book
Storage
bucket profile
Create
Planning views
Planning Version
Assign users to
Planning book
Planning
bucket profile
Resources
The resources are used to define
Capacities of equipment, machines, personnel, means of
transport, warehouses
Resource data are relevant for planning order dates,

taking working time and the capacity of the resources
into account
Resource types:
Bucket Resources
Single activity/ Multi-activity resource
Single Mixed and Multi-Mixed Resource
Page 37
Planning Parameters
Parameters in the resource that are relevant to
scheduling, which the system uses in PP/DS and
SNP.
Used to control in detail how the system schedules
orders to resources.
Examples: Overload, Bottleneck resource, activity
overlap periods, etc.
Page 38
Capacity Models
Shift sequences
Shift sequences
Day number
Validity
Breaks
Shift
factors
Shifts
Quantities / rates
A
Time
Page 39
A B
C
D
Volume
$
Weight
Costs
Production Process Model (PPM)
PPM summarizes
Process
Routing
BOM
PPM supports
Location-dependent (PPM ID) and
location-independent (Plan No.)
Min/max lot sizes for the master material
Operations
Set of different process steps Bucketed time
intervals for each process step
Resources assigned to production steps
Validity periods
Planning Order
Page 40
Production Process Model
Header: Costs, Lot size range

Operations:
Supply
material
Pre-assembly
Activities:
Final
Setup
Produce
Product
Tear down
Relationship
Queue time
Resources
assembly
Inspection
In / Out
Sequence
Resource consumption
Page 41
What is a Quota arrangement ?
Supply
Supply Chain
Chain Model
Model
Liste
60% 40%
APO
Sourcing
Sourcing
Production
Production &&
Purchasing
Purchasing
Requirements,
Requirements,
Inventory
Inventory levels
levels
Page 42
Definition
Forecasts,
Incoming quotas
Forecasts,
Customers
Production lead times
Customers orders
orders
Transportation lead times
Quotas
Production Location
Distribution Center (DC)
O
ut
70%
30%
60%
go
i
Product
ng
uo
ta
s
40%
Quota
60%
Incoming Quotas
Page 43
?
Supplier
40%
Supplier
Planning Process
Demand Planning
Forecast
Supply Network Pl.
Production Planning
Detailed Scheduling
Planned
Orders
Recommended
TransportOrders
Deployment
Transport Load
Builder
Transportation Planning
& Vehicle Scheduling
Page 44
Confirmed
Transport Orders
Transport
Orders
SNP Heuristic
Heuristic is an algorithm that

has pre-defined set of
parameters and workflow to
influence the creation of a
solution
The plan is not necessarily
feasible
Planner must use capacity
leveling to formulate a
feasible plan
Page 45
SNP Run Using Heuristics:

Heuristic is used as part of a repair-based planning
process consisting of the Heuristic, Capacity Leveling,
and Deployment
The Heuristic processes each planning location
sequentially and determines sourcing requirements
Heuristic processing lumps all requirements for a given
material in the location into one requirement for the
period
Heuristic determines the valid sources and quantity
based on pre-defined percentages for each source, then
passes the requirements through the supply chain to
calculate a plan
Page 46
SNP Run Using Heuristics:

Heuristic plans all distribution requirements for all
locations in the distribution network before exploding the
BOM and processing dependent demand in the
production locations
System explodes the BOM only when the Multi-level
Heuristic run option is chosen
Scope of the planning run
Multi-level
Network
Location
Page 47
Factors Considered in the Heuristic Run
Transportation lanes
Lead Times
Quota arrangements
Lot sizing
Scrap
Component
Availability
PPMs
Page 48
Location Products
SNP Demand Profile
SNP Supply Profile
Demand Profile
Heuristic Processing - Results

Replenishment Plan
list of procurement
production orders
transportation orders
Results can be viewed in the interactive planning

table
If the Level ID option is used, the Heuristic calculates
an intersection of the following entries
Model Version
Products and Locations
Level ID from product-location hierarchy
Page 49
Capacity Leveling
Capacity leveling is a function within Interactive
Planning
Enables to smooth production schedule
Manual or using a methods-based approach
Provides opportunity to build up inventory or increase
capacity
Alternatives can be easily analyzed
Re-plan even re-forecast before putting the plan into
production.
Page 50
Optimization-Based Planning
Models
In constraint-based planning, production processes can
be represented as optimization models.
A production model based on optimization consists of
Objective Function(s), Decision Variables, and
constraints based on market conditions, physical
processes, and resources/capacity.
These kinds of models are usually called mathematical
programs.
Page 51
Optimization - Components
Page 52
Decisions variable are the

independent variables of the
problem
The Objective Function is
the single benchmark for
evaluating all combinations
of decisions that satisfy the
constraints
Constraints represent
limitations on which decision
can be made and how
decisions can be made
F(x,y2)=
Optimization of the Network
Sourcing
Sourcing
production
production &&
purchasing
purchasing
requirements
requirements
Control costs
Penalty costs
Forecasts
Forecasts
Customers
Customers orders
orders
Priorities for:
demand types
defined via
costs
Goal: Minimize costs

Goal: Maximize Profits*
Page 53
Optimization Methods
Linear Programming
Continuous Linear Optimization Problems
Primal Simplex Method
Dual Simplex Method
Interior Point Method
Discrete Linear Optimization Problems

Mixed Integer Linear Programming
Page 54
Prioritization
Decomposition
Vertical Aggregated Planning
Horizontal Aggregated Planning
Discretization
Optimization Parameters
Decision Variables
Production lot sizes
Transportation lot sizes
Capacity increase
Objectives
Lateness
Storage costs
Transportation costs
Production costs
Penalty for increasing
capacity
Penalty cost for not
maintaining safety stock*
Penalty cost for late or non
delivery*
Page 55
Constraints
Production capacities
Transportation capacities
Handling capacity
Due dates (demands)
Safety stock
Discrete Values
Production Lot Size
Transportation Lot Size
Optimization Relevant Profiles
SNP Optimization Profile

Specifies the linear programming method to be used and
the constraints to be considered during the Optimization
run
SNP Cost Profile

Specifies the weight given to different categories of costs
in the objective function
Optimization Bound Profile*

Specifies the time buckets where the new plan is
constrained by an upper and lower limit on the allowable
change
Page 56
Optimization Total Costs
Total Cost (Sum Total)
Page 57
Source of cost data
Production
PPM
Storage
Resource
Storage expansion
Resource
Penalty cost for safety stock
Cost Profile
Transport cost
Resource
Handling capacity expansion
Resource
Transport capacity expansion
Resource
Production resource expansion
Resource
Penalty for non-delivery
Master data
Delay
Penalty
Procurement
costs
Master data
Master data
SNP Optimization Run Results
Page 58
Distribution Plan
Production Plan
SNP Resulting Costs
Alerts
CTM Process (Overview)

Constraint-based
heuristics to conduct multisite checks of production/
transportation capabilities
Supply categorization
Demand prioritization
Demand
Prioritization
CTM Engine
Create CTM application
model
Match supply to demand
using the CTM algorithm
Page 59
Supply
Categorization
CTM Process
CTM Engine
Phase 1
Build CTM
application
model
Phase 2
Match supply
to demand
Orders in
liveCache
Demand Prioritization
Demands
forecasts
(from APO Demand
Planning)
sales orders
Prioritization based on
order type
customer priority
product priority
due date
Defined in Sort Profile*
Page 60
Prioritized
demands
Demands
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Supply Categorization
Supplies include
inventory
purchase orders
Categorization is based
on supply limits
for each location
for each product
Supplies
Categorized
Supplies
excess
normal
target
Page 61
CTM Results
Categorized
Supplies
Prioritized
Demands
Capable to Match
1.
2.
3.
4.
5.
6.
7.
8.
9.
Multi-site capacity and

transportation capability check
Page 62
10.
11.
Rule Based Capable to Match*

Allows to influence the supply and demand matching
process, depending on a demands specific attribute
Determines the following based on the attributes of
individual demands
find the product/location substitutes for the particular
demand with substitution
influence the solution process for a particular demand using
demand-dependent constraints
Page 63
Sales and Operations Planning: Overview

Supports interactive SNP Planning
create a feasible plan for the entire supply chain
Considers resource capacities (Production and

Transport)
Flexible Controlling via quotas, priorities,and cost
Parameters are time dependant
No lot size Planning
No optimization
No Orders
Page 64
Sales and Operations Planning:

Overview
Product
DP: (dis-) Aggregation
Location x Product
Location x Product x Channel
SNP: Propagation (finite, multi-level)

Page 65
Sales and Operations Planning:

Features
Pre-configured Planning Environment

Bucketed Planning based on Time Series
Aggregated Planning possible
Fully integrated with Demand Planning
statistical forecasting
promotion planning
Planning books can be configured to compare real

world (OLTP) with tactical plan
Page 66
Conversion of SNP Orders to PP/DS Orders
System automatically stores the final supply

network plan in liveCache
Converting SNP Orders to PP/DS orders
makes it available for finite scheduling
PP/DS enables to synchronize production
planning with execution
PP/DS creates a viable plan
Two ways of conversion
Conversion of SNP Orders in Production Horizon
Conversion of Individual SNP Orders
Page 67
Collaborative Supply Planning Overview

The goal of collaborative supply planning is the
exchange of materials requirements at an early stage
between manufacturers and suppliers so that all
parties involved can adjust their supply and
production plans
Partners can exchange data in two ways:
automatically using time series data exchange between SAP
systems
manually via a web browser for collaboration between SAP
and non-SAP systems
Page 68
Collaborative Supply Planning
Supplier
Both
Manufacturer
Business Benefits
Business Benefits
Better Transparency
Better service level of supplier

Front End
Agreements
Reduced costs by less inventory
Better Service level
More stability in demand
Easy communication with

all related parties
Page 69
Supply
Supply Network
Network
Planning/Production
Planning/Production
Planning
Planning
Exchange
Component
Component
requirements/
requirements/
constr-aint
constr-aint or
or
unconstraint
unconstraint
forecast
forecast
More accurate supply

Supply Network
Planning/Production
Planning
Reduced costs by less inventory

Easy communication with
all related parties
Collaborate on
exceptions
Supply Network
Planning/Production
Planning
Supply Network
Planning/Production
Planning
Collaborative Supply Planning Process Flow
Manufacturer
Supplier
Supply
Planner
Alert
broadcasting
(email)
Da
APO SNP
plan
R/3
Page 70
ta
a
ch
e
ng
orm
f
n
i
e
(al
n
o
ati
ds
Committe
m
rt,e
)
ail
Review
requirements
n
upply pla
replan
Account
Manager
Internet
Supplier
System
Purchase order
Planning Process & Time Buckets
Network Design (ND)
Year, Quarter, Month

Year, Quarter, Month, Week
Quarter, Month, Week, Day
Detailed Scheduling (DS)
Week, Day, Hour, Minute
Deployment
Month, Week, Day, Hour
Transport Load Builder (TLB)

Vehicle Scheduling
Page 71
Week, Day, Hour, Minute
Which products should be planned

using PP/DS?
Externally procured products with long replenishment
lead time
All in-house products produced on a bottleneck
resource
Page 72
What PPDS delivers?

Consistent Model through Production Process Model
Simultaneous Capacity (CRP) and Material (MRP)
Planning
Creation of Feasible Production Plans
Multi-level Forward and Backward Scheduling
Automatic Multi-level Transfer of Changes (e.g.
orders)
Page 73
Why perform Production Planning/

Detailed Scheduling?
Page 74
Improve customer response (due date performance)

Improve throughput
Reduce inventory/Reduce WIP
Reduce overtime expense
Increase asset utilization
Reduce cycle time
How SAP delivers this solution?

Built for quick response (liveCache)
Designed for a multi-plant heterogeneous
environment
Packaged to support R/3 implementations
Simultaneous material and capacity planning
Constraint solving & optimization engine
Simulation and what-if analysis
Exception driven decision support tool
Page 75
Differences between SNP and PPDS
SNP
PPDS
Capacity and product receipts

and requirements are
considered in a bucketed
fashion
No material problem if all
receipts until a certain bucket
equal all requirements until that
bucket
Sequence not relevant
LP-optimizer can:
Both, capacity and

product are checked with
exact time.
It is a potential problem if
material receipt is 1
second after
requirement.
Sequence relevant
Optimizer can:
alter sequence of
existing orders
generate orders
optimize lot size
best for sourcing problems
Page 76
Sourcing,
Confirmation
(SC)
Inventory
Deployment
Transport
Orders
ATP
Actual
Sales
Orders
Historical
DP
Data
Unconstrained SNP
Demand
Plans
Final
Production
Plan
Planned
Orders
PP
Feasible
Schedule
s
Alert Queries
MasterData
SCC
DS
SC
Page 77
Total Process Flow: APO
Core
Order Creation
(OLTP, SNP, PP/DS)
PP
Production Process Model

Evaluation / Selection
MRP
PPM Explosion
Availability check against
unassigned receipts
APO
Order
Stock
Creating Order Network
Transfer
Orders
Page 78
Planning Functionality
PP/DS Setup and Planning Cycle
Page 79
Features in PP
Automatic planning
When integrated with R/3 continually, this can provide the
most up-to-date information
Manual planning
to handle critical products that require particular attention
when planning
Interactive planning
Planning board feature
Cross-plant planning
Stock transfer or between parties in a supply chain
Page 80
Features in PP -
(Continued)
Lot-size calculation
Lot-for-Lot Order Quantity
Fixed Lot Size
Rounding profile
Scrap calculation
Scrap at assembly level
Scrap at activity level
Calculation of Days Supply

Planning with shelf-life data
Page 81
Features in PP
Page 82
(Continued)
Production Process Model (PPM) explosion

Pegging (Fixed and Dynamic)
Evaluation Options
Execution Functions in APO and R/3
Pegging
Demand
Pegging network:
50
Output node
80
50
50
30
20
Order
Page 83
20
Input node
Pegging edge
Used for
supply/demand
allocation
Changes have to be
propagated to all
dependent orders
Find unassigned
order quantity
APO offers fixed and

dynamic pegging
Global setting in
APO
Strategy Profile
& Scheduling Modes
Page 84
Scheduling Strategy Profile
Infinite Scheduling
Schedule an operation without considering resource
capacity
Finite Scheduling
Schedule an operation only when resource capacity is
available
Alert is only generated when planning using finite
scheduling only
Finite
Infinite
Strategy
Strategy
Finite Scheduling
Infinite Scheduling
Page 85
Finite scheduling
Infinite scheduling
Alert display
Alert display
Infinite scheduling
No Alert display
Infinite scheduling
No Alert display
Scheduling Modes
Infinite loading
Finite loading only
forward
Finite loading only
backwards
Finite loading with
direction switch
Search for a slot in
schedule
Page 86
Insert operation to
close gaps in
schedule
Squeeze-in
operation
Add an operation at
end
Dispatch to nonworking time
Finite Scheduling - Backward Only
Available
New Order
Occupied
Customer Order
Desired date
and quantity
Resource 1
Resource 2
Resource 3
today
Page 87
confirm
Finite Scheduling - Forward Only
Customer Order
Available
Occupied
Desired date
and quantity
New Order
Resource 1
Resource 2
Resource 3
today
Page 88
Confirmed date
and quantity
Finite Scheduling - Backward with

Reverse
Available
1st Loading attempt
Occupied
Final loading
Customer Order
Desired date
and quantity
Confirmed date
and quantity
Resource 1
Resource 2
Resource 3
today
Page 89
Simultaneous Material and Resource

Planning
Available
Occupied
1st loading
Infinite planning strategy
Due Date
Resource 1
Resource 2 (bottleneck)
Resource 3
ALERTS:
Material 1
Resource Overload
Supplier Delivery Time Violated
Delivery time
today
Page 90
Simultaneous Material and Resource

Planning
Available
Occupied
1st loading
Infinite planning strategy
2nd Loading
Finite planning strategy
Due Date
Feasible Due Date
Resource 1 (finite scheduling)

Resource 2 (bottleneck)
Resource 3
Material 1
Delivery time
today
t
Page 91
Features in Detailed Scheduling

Determine production date by taking account of the
constraints entered in the strategy profile and in the
resources for scheduling
Changes in DS (e.g. orders, operations, ) will
automatically propagate through to all relevant
objects
Taking sequence-dependent set-up times and/or
costs into consideration
If a constraint is relaxed (this is not considered in
optimization run) and it is violated in scheduling, the
system creates alerts
Page 92
Real Time Planning vs Batch

Optimization
Real Time Planning
supports interactive scheduling
finds a feasible solution
Batch Optimization
takes into account
complete situations
real time answering
optimizes feasible
solutions
Examples
answering time depends

on user-defined run time
backward/forward propagation
scheduling in free slots
simultaneous material and
capacity planning
Page 93
Examples
complete rescheduling
of planning window
Concept: Optimization
Order 1
Order 2
Order 3
Order 4
Resource 6
Resource 5
Resource 4
Resource 3
Resource 2
Resource 1
time
Page 94
Concept: Optimization
Order 3
Order 4
Order 1
Order 2
Resource 6
Resource 5
Resource 4
Resource 3
Resource 2
Resource 1
time
Page 95
Foundation of Optimizer in DS:

Metaheuristics
Objective
obtain quality solution for given time frame
(scalability for a given problem size)
Purpose
simplify the problem
Metaheuristics consists of
time decomposition
resource decomposition
constraint relaxation
Local improvement Strategy
Objects
(resources, orders,
ops, constraints, )
Reduced
opt model
Focus on a sub-problem and optimize

planning window
Time
Page 96
Metaheuristics Time Decomposition

Resources
Current window
Time
Sliding window (Rolling time Horizon)

1. Optimize only in window
2. Move window by a time delta
3. Go to first step
Page 97
Metaheuristics Bottleneck Decomposition

Resources
Bottleneck
Time
Bottleneck
1. Determine bottleneck
2. Schedule bottleneck resources only
3. Fix sequence on bottleneck resource
4. Schedule all resources
Page 98
Metaheuristics Constraint Relaxation

A way to make a constraint less restrictive or
remove certain constraints
5 choices
Page 99
Remove max time constraint

Set resource utilization rate to 100%
Remove calendar
Do not consider sequence-dependent setup
Undo fixing of orders/operations/activities
Scope and Size of Optimization: An

Example
Non-work times
Fixed operations
Operations, that
may be rescheduled
Relationships
Optimization range
Page 100
Start
End
Transferred Resources B, C, D
Optimization Model in PPDS:

Scheduling
Decision Variables
Resource Allocation
Alternative Machines
Alternative Storage
Start dates
End dates
Objective Functions (Minimize)
Total Lateness
Maximum Lateness
Total Leadtime
Total Setup Times
Total Setup Costs
Constraints
Time Constraints
Maximal (Shelf Life)
Minimal
Deadlines
Production and Storage
Capacities
Calendar (Shifts and Breaks)
Sequence- and Resourcedependent Setup times
Resource Network
Breakable activities
Effectivity of BOMs and
Routings
Productivity (per Shift)
Page 101
Two Ways to Perform DS

Optimization Procedures
Constraint Programming
Constraint propagation
Branch and bound
Genetic Algorithm
Priority rules
Sequencing
Page 102
Constraint-Based Programming
Approach/method
checks hard constraints during scheduling
propagates constraints during scheduling
additional reduction of the search space
early detection of dead ends
reduces back tracking
Tradeoff
Dynamic propagation needs time
but improves quality of search decisions
Advantages
High-performance constraint propagator (iLog)
Dedicated to complex scheduling problems
Example: shelf life / expiration
Page 103
Constraint-based Programming:
Example
Propagate consequence
of each decision
Dynamic constraint
propagation
checks hard constraints
during scheduling
Prune the search tree

Remove unfeasible
solution
Remove worse solution
Page 104
Initial Solution
Solution
with
Changed
Variables
Genetic Algorithms
(GA)
Based on Concept of Evolution

Population of candidate solutions
New candidate solutions by
Crossover/Recombination
Mutation
Exchanging sequence of activities
Change resource allocation of activities
Page 105
Preferring the better ones as parents

Eliminating the worse ones
GA - Procedures
Generate initial schedules
Selection of good schedules

Generate new schedules by
mutation and recombination
Evaluate new schedules
Survival of new schedules
Page 106
GA- Advantages
High-performance in sequencing
Dedicated to not too complex scheduling
problems
Feasibility should be not the problem
Example: no shelf life / expiration
Page 107
Optimization Profile
Maintain in
Customizing
Settings
Optimizing procedure
Constraint-based
Scheduling
Genetic Algorithm
Runtime
Objective Functions and its
Weights
Page 108
Total lead-time
Set-up times
Set-up costs
Maximum delay
Average delay
Constraint Relaxation
Remove maximum time
constraints
Set utilization rate of all
resources to 100%
Remove calendar
Do not consider set-up
times/costs
Undo fixing of activities
Techniques in Solving
Complex Production
Processes and
Optimization Models in
APO
Page 109
APO Computational Solvers
APO Solvers
A wide variety of computational solvers

applied to specific planning functions with
industry-specific variations
Linear Programming /
Mixed Integer Linear
Programming
Supply
Network
Planning
Genetic Algorithms,
Constraint-based
Production
Programming
Planning &
Scheduling
Heuristic Methods
Exponential Smoothing
Holt Winters
Multiple Linear Regression
Page 110
Demand
Planning
Complexity in Real-life Production

Models
They are too complex to solve. For example, we may
have nonlinear equation(s), integer decision
variables, scale, ...
In such cases, we will have to rely on algorithm,
heuristics, and other intelligent methods.
Most APS systems mix optimization and heuristic
methods.
Page 111
Where is Constraint-based Planning

Used in APO?
Creating Production Process Models (PPMs) in SNP
and PP/DS are based on Theory of Constraint and
Constraint-based Planning
Where are the decision variables? Where are the
constraints identified? Where are the objective
identified?
For example, in PPDS, the decision variables are
given in the planning table (which resource is used to
produce a given order and its start and end times);
objective function is given in the optimizer screen;
and the constraints are stated in master data.
Page 112
Example: Decision Variables in DS
End Time
Start Time
Page 113
Example: Objective Functions in DS
Optimization methods
Obj. functions and its weights

Page 114
Optimization Models in APO
Page 115

SAP APO Overview

Uploaded by

Copyright:

Available Formats

SAP APO Overview

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SAP APO Overview

Uploaded by

Copyright:

Available Formats

What are the main components of SAP APO?

What are the main components of SAP APO?

How does liveCache help improve performance in SAP APO?

How does liveCache help improve performance in SAP APO?

SAP APO Overview

MindTree Consulting Proprietary & Confidentia

SAP APO Architecture

SAP Advanced Planner & Optimizer (APO)

Application Link Enabling

Model Generator, Mapping, Connectivity

1997 SAP AG APO 2

MindTree Consulting Confidential

SAP APO Application Architecture

Supply Chain Cockpit

MindTree Consulting Confidential

Multiple System Environment

MindTree Consulting Confidential

Optimized for realtime scheduling and

Integration with R/3

Standard Transaction Handling

MindTree Consulting Confidential

MindTree Consulting Confidential

Main objective = higher performance:

MindTree Consulting Confidential

liveCache Integration with R/3

Optimizer routines are developed in C++, which increases

MindTree Consulting Confidential

SAP APO Integration

APO - R/3 Integration

Core Interface Facility (CIF)

MindTree Consulting Confidential

CIF : Core Interface

MindTree Consulting Confidential

CIF Functions Architecture

APO -> ERP

MindTree Consulting Confidential

Modules being Implemented

MindTree Consulting Confidential

MindTree Consulting Confidential

Supply Chain Planning Cycle

Supply Network Planning (SNP)

Production Planning (PP)

Transportation Plan / Vehicle Sched.

MindTree Consulting Confidential

Planning Area Functionality

MindTree Consulting Confidential

Planning Area Flow

Assign Key Figures

APO Planning Version

MindTree Consulting Confidential

MindTree Consulting Confidential

MindTree Consulting Confidential

MindTree Consulting Confidential

SAP APO Overview:

MindTree Consulting Proprietary & Confidentia

The Supply Chain: Original Supply

Supply Chain Optimization

Increase Customer Responsiveness at Least Cost

MindTree Consulting Confidential

What is Supply Network