What is Activity Profiling?

1
What is Activity Profiling?
Warehouse Activity Profiling is the analysis of
historical sales transaction data for the
purposes of projecting warehouse activity and
determining storage mode, physical layout, work
flow processes, and labor and equipment
requirements.
2
Profiling Database Source Data
ITEM MASTER

Inventory Snapshots Average Inventory Levels
Order Header Order Detail
SKU Number Description Item Cube Pieces Per
Case Cases Per Pallet Division Product Group Item
Weight
Item OrderedQtyUnit of Measure
3
Developing Profiling Reports Graphs
InventoryMasterData
OrderData
STEP 1 CONSOLIDATE CALCULATE
ItemMasterData
4
How Do You Design a Warehouse?

• Two Ways To Design a Warehouse
• Storage Driven Approach via Cube Analysis
• Picking Driven Approach via Order Analysis

Storage Driven
Picking Driven
5
What is the Storage Driven Approach to Design?
Define Your Storage Zones Design Your Forward
Pick Areas Define How You Will Plan Pick Orders

• PART I
• PART II
• PART III

6
Designing a Warehouse
Part I
Define Your Storage Zones
7
Categorize Items By Cubic Ft of Inventory
Calculate the cubic feet of storage that each
item requires and then assign it to an inventory
container of the appropriate size.
Pallet Rack
Bin Shelving
Multi-PalletDrive In Rack
Drawers
.125
1.5
40.0
320.0
Cubic Feet of Storage Required For An Item
8
Develop an Inventory Container Graph
Now you can begin to think about what storage
modes might be reasonable candidates for the
merchandise you are storing
9
Develop a Pick Size Classification Scheme
Next develop a classification scheme for picks
based on the size of the pick. Usually designers
will use pallet, case, and piece pick sizes
Piece Pick
Case Pick
Pallet Pick
10
Assess the Activity In Each Inventory Container
Assess the activity in the larger containers to
see if there is the possibility that some of the
items should be moved to a forward pick area.
The decision will be driven by the of such
picks in the container and the overall size of
the larger container storage area.
11
Designing a Warehouse
Reserve Areas
Part II
Define Your Forward Pick Areas
Forward Pick Areas
12
General Process for Forward Pick Design

• Questions that Must Be Answered About the Forward
  Pick Area(s)
• How many forward pick areas do you need?
• Determine how many SKUs should go on the pick
  line
• Removing unusual SKUs from the pick line
• Sequence the SKUs on each pick line

13
You will likely have multiple forward pick areas
For each Pick Size you need to decide if there
are a lot of picks associated with a relatively
small subset of the items. If so, you will
likely want to set up a forward pick area for
that Pick Size.
14
Determining How Many Items in Forward Pick
Generally to determine how many items you are
going to put in the forward pick area you look at
the tradeoff between adding an item into the
forward pick area and the of additional orders
you are then able to complete in that area.
Case
Trade Off Space Utilization and Efficiency
Number of
Days
Picks
SKUs
Picked
Filled
20
33
99
30
46
97
40
52
95
50
67
93
60
73
92
70
79
90
80
81
83
90
84
79
100
88
68
110
92
63
120
92
52
130
95
44
140
98
33
Number of SKUs
150
98
25
160
100
22
15
Determining How Many Items in Forward Pick
90
16
Designing a Warehouse
Part III
Define How To Plan Pick Orders
17
Wave Planning Picking Approaches
Daily Order Pool
Orders of this type get released to the floor
and picked in the following manner every X hours
Orders of this type get released to the floor
and picked in the following manner every Y hours
While designers make assumptions at the start of
a design about how the bulk of the orders will be
released and picked, the details behind their
thinking are not usually flushed out until the
end of the project. They often also wait until
the end to define the planning and picking
approaches for the exceptional orders.
18
Ways in which you can process orders differently

• Order Selection Criteria Groups
• Rush vs Regular Orders
• Geography (West Coast vs East Coast)
• Orders Requiring Personalized Merchandise
• Single vs Multi-Line Orders
• Types of Picks Needed to Complete Order
• Order Cube (Sm Pkg vs LTL vs TL)

19
Assess the Significance of Single Unit Orders
Units/Order as a Percentage of Total Orders
45 of all Orders are single unit orders.
One of the greatest opportunities to improve
warehouse efficiency is choosing a different
mechanism for picking single unit orders from
multi-unit orders.
20
Assess the significance of grouping by area
Reserve(Pallet)Area
Orders Completed By Area
Case Forward PickArea
Piece Forward PickArea
Orders that require merchandise coming from
different storage areas within the warehouse may
need to be picked differently.
21
Deciding on a Picking Approach
Single Order Picking
Multi-Order Picking
Batch Picking
After the different groups of orders have been
identified, the designer has to make a decision
about how each group of orders will be picked.
22
How Will Orders in Forward Pick Be Picked?
23
Deciding on a Picking Medium
Label Picking
Radio Frequency (RF) Barcode Picking
Voice Picking
For each picking approach you need to decide on a
mechanism for how picks will be communicated to
pickers.
Pick To Light
24
How Do You Plan Pick Different Orders
Order Group
Pick Method
Pick Medium
Small Cube Multi-Line Orders
Multi-Order Picking
RF Terminals
Single Line Orders
Batch Picking
Labels
25
Summary of Warehouse Design Process
Define Your Storage Zones Design Your Forward
Pick Areas Define How You Will Plan Pick Orders

• PART I
• PART II
• PART III

26
Observations

• Every descriptive tool or technique seems to be
  based on a specific need
• Profiling/design is less about describing an
  as is warehouse, than about saying how it
  should have been
• Its hard to integrate the different descriptive
  tools and techniques
• Can we build a comprehensive, computational
  description from which all the different
  needs can be met?

27
Reference Model
Domain
Model
Instance
28
AMPL, AIMS, GAMS, and other modeling languages
incorporate a reference model for the domain of
optimization models, and are used to create
instances of optimization models.
29
Can reference models be developed for the domain
of discrete event logistics systems, or for
subsets of the domain, e.g., warehouses,
factories, and supply chains?