JADE Intro Case PowerPoint - Supply Chain Network Design

Report
Chapter 9:
JADE Intro Case
NetworkDesignBook.com
Material by Michael Watson, Sara Lewis, Jay Jayaraman, and Pete Cacioppi, 2012
Material Needed

Files needed:

Zipped file called: JADE
Intro Case.zip

LNP XE Version 7.2.24 is
the latest

If you click on the symbol
in the upper left and click
options, you can change
the Excel settings
Material by Watson, Lewis, Jayaraman, and Cacioppi
Opening the model

Unzip the file contents into a new directory

Open LogicNet Plus XE (LNP XE)

Click on the Open link and go find the directory you just
installed the files into

Project should open
Material by Watson, Lewis, Jayaraman, and Cacioppi
JADE Paint and Coverings Background–
Corporate Background




JADE Started Business in 1906 from the Nassau-Suffolk, NY area. They still have their
main distribution center there (red triangle on the map)
They originally started just making paint, but now sell specialty paints, wood treatments,
ready-mixed concrete for patching, and other such products
In their markets, they have a strong brands that command a premium and loyalty from the
professional contractors
Within the last ten years, through acquisitions and strong execution, they have broken in
the West Coast and greater Texas markets
 Although these new markets helped
grow revenue, they were not good
from a transportation point of view. A
new DC was put up in Phoenix to help
handle these markets
 The last VP of the Supply Chain just
left the firm and the new one is under
intense pressures to take costs out of
the supply chain. It is not clear the
company can stay in business with its
current supply chain
Material by Watson, Lewis, Jayaraman, and Cacioppi
JADE’s Supply Chain




JADE has kept its east coast DC in the Nassau-Suffolk, NY area– although the east coast
market extends into Florida and the Midwest
JADE sells four basic product families. Currently, each of the plants is set up to make just
one of the product families.
JADE has three plants in the US and one plant in China. All China product comes into the
port at Long Beach.
JADE’s manufacturing plants are known in the industry for their low cost and high quality
 The plant in China was a bit of an
experiment and the result of an
acquisition. It is unusual for such a
heavy product to be made so far away
from its market. Competitors raised
an eyebrow at this decision
 Their plants and DC’s are set up to
handle rail shipments, so the
shipments from the plants to the
warehouses are a mix of truck and rail
Material by Watson, Lewis, Jayaraman, and Cacioppi
JADE Case Study Objectives

Use LNP XE to help JADE reduce costs in their supply chain

Learn how to use LNP XE
 Learn basic navigation
 Learn the core set of key files to get models up and running
 Learn how this model was built
 Learn how to run and analyze scenarios

This model has some simplifications. Our primary objective, right now, is to
learn LNP XE. We will add sophistication to the case as we progress through
the exercises
Material by Watson, Lewis, Jayaraman, and Cacioppi
Mapping JADE’s Supply Chain to LNP XE
Manufacturers
Assembly Plants
Vendors
Port of Entry
Distribution Centers
Plant Direct Shipment
Cross Dock Facility
Temporarily Leased Space
End Customers
Retailers
Demand Points
Product Destinations
LNP XE will minimize the cost to get product from the plants to the
customers, considering all business rules and choices you give it.
Material by Watson, Lewis, Jayaraman, and Cacioppi
When you are mapping a supply chain to LNP, you want to think
about the following questions. Your answers to these will dictate
how you set up the model, what data you will need to collect, and
the type of results you’ll see

What Costs do you want the model to consider when it runs?:
 Most times it is to minimize cost and occasionally to maximize profits
 You need to decide what costs to include

What are your Business Rules, Constraints, or other Limits?
 When you minimize cost, a constraint is that you need to meet all demand
 You also define limits or constraints in terms of what can be made where,
how much can be made at each location, warehouse capacity, rules on
which customers can serve other

What Choices or Decisions do you want the model consider?
 When the model runs, what choices do you want to give it. Can you make
Material by Watson, Lewis, Jayaraman, and Cacioppi
Typical Methodology
Kickoff & Data
Collection
Data Analysis &
Validation
Baseline
Modeling
Analyze
Alternative
Scenarios
Final Summary and
Recommendations

In the kickoff, you will spend time mapping your supply chain to LNP XE and determining data that is
required

In a typical project, your first model will be a baseline model. The baseline model is a model of your
supply chain with last year’s data that has all the decisions fixed. This model is then used to compare
against your financial records to make sure the model is running correctly.

In addition to the baseline model, you will also want to run an optimized baseline model. This is how
your supply chain should have performed given your current structure and rules.

The difference in cost between the baseline and optimal baseline is from unavoidable problems or
from execution issues. Often, you can find significant savings jus with this exercise

It is always best to compare your scenarios to the optimal baseline to get a good sense of savings
potential
Material by Watson, Lewis, Jayaraman, and Cacioppi
JADE’s Kick-Off Results:
JADE’s team is going to start the optimal baseline

What Costs do you want the model to consider when it runs?:



What are your Business Rules, Constraints, or other Limits?




We will consider the transportation costs from the plant to the warehouse
and from the warehouse to the customer. For the imported product from
China, we will only consider the cost from Long Beach to the warehouses
For this initial model, we will not consider production costs, warehousing
costs, inventory costs, etc.
For the optimal baseline, we will not allow the plants to change, the plant’s
mix of product to change, or the warehouse locations to change
All product has to move from a plant to a warehouse and then to the
customer. The customers require a mix of product and JADE is not set up
for plant direct shipments
All customers should get all products from a single warehouse
What Choices or Decisions do you want the model consider?

For the optimal baseline, we are allowing the model to determine which
customer should be served from the two warehouses. As a side benefit,
this will determine the total tons moving on each of the lanes
Material by Watson, Lewis, Jayaraman, and Cacioppi
Here is a minimum set of files to build a model.* These are the core
files you will need to get any model built. Within in the forms and
with other forms, you can add as much sophistication as you like
Sites & Products
Production
Info
Plant
ID
Product
ID
WH
ID
Demand
Customer
ID
* If you
haveby
plants
shipping
to customers,
even get rid of the warehouses
Material
Watson,
Lewis,
Jayaraman,you
andcan
Cacioppi
Transportation
Plant to
WH Lanes
WH to
Plant Lanes
WH to WH
Lanes
Plant to
Plant Lanes
Carriers
WH to
Customer
Lanes
Let’s See How JADE’s Data was loaded into
LNP


Open LNP XE
To Open JADE Project




Either click on “Open” in the middle right, click
the Open Icon (the standard folder) in the
upper left
Go to the folder where you unzipped the file
and click on the project.ilnp file
To get to the input data, you need to
either go to the ribbon at the top or the
Navigation panel to the left. When you
click on Data, you see the menu of data
input elements.
We will focus on the six major elements
shown in the ribbon
Material by Watson, Lewis, Jayaraman, and Cacioppi
Scenario Preferences


Click Data
Ribbon, go to
Preferences and
select Scenario
Preferences
For JADE, we are
going to keep
things simple and
set up everything
in terms of total
tons. So, all units
of measure will
just be in tons
Material by Watson, Lewis, Jayaraman, and Cacioppi
Products


Click on Products in the Ribbon or Product Details in the
Navigation panel
Here are the four product families for JADE. Note that all
products are active. Since we are only using tons for units
of measure these fields are set to 1.00.
Material by Watson, Lewis, Jayaraman, and Cacioppi
Customers



Click on Customers in the Ribbon or Customer Details in the Navigation panel
These are the market areas that JADE services
Note that that we see a constraint here--- “Single Source” is checked for these
customers. This ensures that all customers get all their product from a single
warehouse
Material by Watson, Lewis, Jayaraman, and Cacioppi
Demand


In the Navigation panel, under Customers, click on Demand
This shows the total demand for each of the four product
families at each customer. This is JADE’s annual demand
from last year
Material by Watson, Lewis, Jayaraman, and Cacioppi
Demand– Some Facts

If you click on the Graph Icon
at the end of the row of Icons
on the demand form, you can
get some additional insight
into the demand numbers.

First, we’ll select “By Product”
under the Pareto Chart Type.

Next, we’ll come back and
select “By Customer” and
show the top 15 customers
Material by Watson, Lewis, Jayaraman, and Cacioppi
This shows JADE’s breakdown in volume by product family.
Click on “Graph Options” to go back and select another view
Material by Watson, Lewis, Jayaraman, and Cacioppi
Here are the Top 15 Customers
Material by Watson, Lewis, Jayaraman, and Cacioppi
Warehouses



Click on Warehouses in the ribbon or Warehouse Details in the Navigation panel Under Warehouses
Here is a list of the existing two warehouses, along with some others JADE wants to consider
In the first model, they want the model to use only the two existing warehouses. So, you can see that no
warehouse is active, except Phoenix and New York. Note that those are also set to “Fixed.” This tells
the model that it has to have these warehouses in the solution.
Material by Watson, Lewis, Jayaraman, and Cacioppi
Warehouse Capacity



In the Warehouse Navigation Panel, click Warehouse Capacity
By default, once a warehouse is set up, it has infinite capacity and no
costs
For JADE, we are not including these cost numbers in the initial
analysis, so we can leave this as it is
Material by Watson, Lewis, Jayaraman, and Cacioppi
Plants




Click on Plants in the ribbon or Plants Details in the Navigation panel Under
Plants
We use the term “Plant” generically. It refers to the sources of products. So, the
plants could be owned by the companies, could be suppliers, or in some cases
(like for JADE), could represent the port of entry.
Initially, JADE is not interested in changing plant locations. So, all four of these
plants are fixed
When you set up a Plant, you can specify its type as either “Supplier” or “Mfg
Plant.” The distinction only determines how much detail you want to model
inside the plant. Supplier type requires a lot fewer details. JADE was not
interested in doing detailed plant modeling, so we set the type to “Supplier.”
Material by Watson, Lewis, Jayaraman, and Cacioppi
Supply Details


In the Plant’s Navigation Panel, go to Supply Details
This form shows us the constraints on what can be made where. By setting up the max supply as
greater than 0, it shows where a product can be made. A “0” indicates that this plant cannot make the
product. Note that JADE is not yet modeling production costs.
Material by Watson, Lewis, Jayaraman, and Cacioppi
We’ve now set up all the “Sites & Products” for
JADE
Sites & Products
Production
Info
Plant
ID
Product
ID
WH
ID
Demand
Customer
ID
* If you
haveby
plants
shipping
to customers,
even get rid of the warehouses
Material
Watson,
Lewis,
Jayaraman,you
andcan
Cacioppi
Transportation
Plant to
WH Lanes
WH to
Plant Lanes
WH to WH
Lanes
Plant to
Plant Lanes
Carriers
WH to
Customer
Lanes
Lane Visualization– Showing the Possible Paths Through the
Supply Chain

Click on Lanes in the Ribbon and then the

In JADE’s case, we have a simple set up. All plants can ship to all warehouses. And, all warehouses can ship to all
customers. If you do not see a lane type, that move is not possible. So, there are no lanes from plants to customers
since JADE cannot do this
Note that the groups “All Plants, All Warehouses, and All Customers” are default groups . When the model runs, LNP XE
will automatically calculate the unique Origin/Destination pairs for individual sites. You can also create new groups to
build new rules.
In the details in the bottom right, you can see which carrier is assigned to each of the lanes.


Material by Watson, Lewis, Jayaraman, and Cacioppi
Carriers--- Rate Structures



Click on Data | Transportation | Carriers
To start, JADE used high level estimates for their
transportation costs. They used a cost of $0.07 per
ton/mile for moves from plants to warehouses and $0.12
per ton/mile for moves from the warehouses to
customers.
Since some customers were very close to warehouses,
they put in a min charge to capture the short move.
Material by Watson, Lewis, Jayaraman, and Cacioppi
We’ve now have a complete model for JADE
Sites & Products
Production
Info
Plant
ID
Product
ID
WH
ID
Demand
Customer
ID
* If you
haveby
plants
shipping
to customers,
even get rid of the warehouses
Material
Watson,
Lewis,
Jayaraman,you
andcan
Cacioppi
Transportation
Plant to
WH Lanes
WH to
Plant Lanes
WH to WH
Lanes
Plant to
Plant Lanes
Carriers
WH to
Customer
Lanes
Now, Let’s Run JADE’s Baseline



Go to the
Optimize Tab
and click on
Parameters
Keep all the
default settings
Click “Run” on
the bottom right
Material by Watson, Lewis, Jayaraman, and Cacioppi
High-Level Analysis–
Go to Solutions Ribbon and click Compare Solutions to see the total cost
Go to the Navigation Panel and click on Customer Service, then by Distance, then by
the 2nd Tab
Material by Watson, Lewis, Jayaraman, and Cacioppi
Summary Reports



In the Solutions Ribbon, click on Summary Reports, 1st tab
This has some information you saw on the Comparison report
It also provides the average distance for the different types of moves
Material by Watson, Lewis, Jayaraman, and Cacioppi
Warehouse Summary Report



Go over to the 2nd Tab
This shows the total tons (Throughput) going through each
DC. New York handles about 2X the volume of Phoenix
It also shows the weighted distance to the customer and the
inbound and outbound transportation costs
Material by Watson, Lewis, Jayaraman, and Cacioppi
Plant Details


In the Solutions Ribbon, click on Plant Details
This reports shows which plant made which product and
how much it made
Material by Watson, Lewis, Jayaraman, and Cacioppi
Transportation Reports


Go to Solutions in the Navigation Panel and click on
Transportation Report
This report gives you detailed information for each of the
individual lanes– (each line on the map).
Material by Watson, Lewis, Jayaraman, and Cacioppi
Landed Cost Report


Go to Solutions in the Navigation Panel and click on Trans | Standard
Reports | Landed Cost Output
This report shows the landed cost to get a product to a customer
Material by Watson, Lewis, Jayaraman, and Cacioppi
Now, Let’s Try to Improve JADE’s Supply Chain:
The most obvious question the new VP of Supply Raised was whether the 2
warehouses were in the right location and whether 2 was the right number.
Here is how we now want to set up the model….
 What Costs do you want the model to consider when it runs?:
– We will consider the transportation costs from the plant to the warehouse and from the
warehouse to the customer. For the imported product from China, we will only consider
the cost from Long Beach to the warehouses
– For this initial model, we will not consider production costs, warehousing costs, inventory
costs, etc.
 What are your Business Rules, Constraints, or other Limits?
– For the optimal baseline, we will not allow the plants to change, the plant’s mix of product
to change, or the warehouse locations to change (this last part is no longer true)
– All product has to move from a plant to a warehouse and then to the customer. The
customers require a mix of product and JADE is not set up for plant direct shipments
– All customers should get all products from a single warehouse
 What Choices or Decisions do you want the model consider?
– For the optimal models, we are allowing the model to determine which customer should
be served from the two warehouses. As a side benefit, this will determine the total tons
moving on each of the lanes
– We now want the model to determine the optimal two, three, four, and five warehouses
Material by Watson, Lewis, Jayaraman, and Cacioppi
Running the Optimal Two Warehouses


Right click on the scenario and
create a copy of the scenario
Right click on the copy and rename
the scenario


S2 Optimal 2 WH
Right click on the new scenario and
click “Set as Active”

NOTE: It is important that you make sure
the red checkmark is on the scenario you
want to work on
Material by Watson, Lewis, Jayaraman, and Cacioppi
Running the Optimal Two Warehouses: Giving the Model
Warehouses to Choose From









Go to Warehouses
Remember, in the previous run, only the existing warehouses were
active and set to fixed
We now want to open up all 25 warehouses for consideration
Click on the Excel Button in Warehouse Details Form (upper left, 3rd
button from the left). This will take this form out to Excel for editing
Once in Excel, set the active column to TRUE for all 25 warehouses
and copy the status of “Potential” over the two warehouses set to
“Fixed.”
Save and close the Excel file
Back in LNP XE, click “OK” to accept your changes
In the Warehouse form, we are now giving the model choices on
warehouses
REMEMBER: We already set up our transportation so that the model
knows how to get from All Plants to All Warehouses and from All
Warehouses to All Customers.
Material by Watson, Lewis, Jayaraman, and Cacioppi
Running the Optimal Two Warehouses: Telling
it to Pick Just Two





In the Navigation Panel, go to Constraints, and Group Constraints
Click on New and then name the constraint “WH Limit” have it apply to warehouses, and “All
Warehouses”
Then set a min and max of 2 warehouses
Close this form and back on the Parameters form, hit “Run”
Analyze the same set of reports we did previously to understand the solution
Material by Watson, Lewis, Jayaraman, and Cacioppi
Optimal 2 DC’s
In the customer service report, click on
Advanced, and do a Side by Side comparison
with the baseline
Material by Watson, Lewis, Jayaraman, and Cacioppi
Questions on the Optimal 2 DC’s

What is different between this scenario and the baseline scenario?

How did the mix in volume change between the east and west DC
change?

How did the average distance change between the scenarios?

What didn’t change between the scenarios?

Where did the savings come from?
Material by Watson, Lewis, Jayaraman, and Cacioppi
Running the Best 3, 4, and 5 DC’s


Right click on scenario “S2” and make three copies.
Make sure you copy this scenario and not the baseline
Rename them:




In turn, do the following





S3 Opt 3 WH
S4 Opt 4 WH
S5 Opt 5 WH
Set the scenario to active
Go to Optimize | Parameters | Min/Max Group of Sites
Change the Min/Max warehouses to 3 and 3, 4 and 4, and 5 and 5 (HINT–
change the Max first)
Don’t run yet. Once you made all the changes, don’t forget to hit Save
Go to Optimize | Batch Run

Select the three new scenarios and click to run the batch
Material by Watson, Lewis, Jayaraman, and Cacioppi
A Side Analysis on the Optimal Two DC
Solution

Someone on the JADE team pointed out that in the 2 DC
solution the move from New York to Columbus seemed
much more significant than the one from Phoenix to Las
Vegas. So, now the team wants to run another scenario
with Phoenix instead of Las Vegas

Make a copy of S2, rename the copy “S2a Columbus and
Phoenix,” and make this new scenario active.

Go to the Warehouse form

For the Status of Phoenix and Columbus, set to Fixed

Now, run this scenario

Since this is a copy of S2, the model can only pick 2 location. Since we
fixed both Phoenix and Columbus, those will be the 2 that the model
chooses
Material by Watson, Lewis, Jayaraman, and Cacioppi
Comparison of all the Solutions
Material by Watson, Lewis, Jayaraman, and Cacioppi
Case Study Questions

What new locations did the model pick as it from 2 to 5 optimal DC’s

How much impact did the change from Phoenix to Las Vegas have?
Why?

What is the rate of change in the savings?

In the 5 DC solution, even though the Columbus DC is very close to
the plants, it still chooses to serve customers from the additional
DC’s, why? In other words, why not save on inbound transportation
and serve most customers from Columbus?

Even though we don’t have the cost of opening a warehouse in the
model, what information does the model give us in terms of what this
cost needs to be for these solutions to be effective?
Material by Watson, Lewis, Jayaraman, and Cacioppi
Independent Exercises

What are the optimal 3 warehouses and cost if all product
comes from China (enters in Long Beach)?

Hint– to make all product in China, you want to go to the Supply Details
and set the Max Supply to zero for all plants but the one for China. For
this plant, make sure the max supply allows it to make all four products

What would have been the cost if they moved all
production to China, but kept the original 3 warehouses
from scenario S3?

What is the value if all plants could make all 4 products?
How much better of a solution could you derive?
Material by Watson, Lewis, Jayaraman, and Cacioppi

similar documents