inventory optimization in a retail multi echelon …...average inventory (unit) supplier store dc...

Inventory Optimization in a Retail Inventory Optimization in a Retail MultiMulti--Echelon EnvironmentEchelon Environment

Rintiya Arkaresvimun

22 MAY 2008

Advisor: Larry Lapide

Arkaresvimun 06 May 2008 2

Agenda

Introduction to RetailCo’s Supply ChainResearch QuestionsInventory model developmentOptimal result and sensitivity analysis


RetailCo Supply Chain Network

(R,S)Weekly review, Order-up-

to-level with case pack quantity

(R,S)Daily review, Order-up-to-

level with case pack quantity

.....

RetailCo’sDC

SupplierCo’s Warehouse

Store 1

Store 2

Store 3

Store 100

(R,S)Period review (Every 3 days or 7 days), Order-up-to-level

This study focuses on 3 echelons : SupplierCo’s warehouse, RetailCo’s DC, and RetailCo’s storesAll echelons’ use periodic, order-up-to-level (OUTL) policiesReview periods are based on fixed schedule basis


Current SituationSingle-echelon approach is used to set up an echelon’s inventory policies

Supplier

Production

Warehouse

Distribution Center

StoresStore Inventory Optimization

DC Inventory Optimization

Supplier Inventory Optimization

What’re the problems?

Suboptimal inventory allocation

Customer service failure

Supplier

Production

Warehouse

Distribution Center

Stores

Multi-Echelon Optimization


“ How should inventory be distributed to optimize the total network carrying cost while still achieving the store target service level ? ”

Additional Questions :

How does an echelon’s inventory policy impact the other echelons?

How does store target service level impact inventory distribution and the other echelons’ service levels ?

How does supplier’s service disruption impact the echelons’service levels?

Research Questions


Inventory Model Development

2-Year Random Daily Store Demand(Poisson)

.....

Store- Segment 1

Store- Segment 2

Store- Segment 3

Store- Segment12

100 Stores are segmented into 12

Store-segments

RetailCo’sDC


Periodic Review, OUTL



OrdersOrders

Delivered Quantity

OrdersDelivered Quantity

Orders

Store Service Level = Number of Day In-StockSupplier and DC Service Level = Fulfilled Demand


Objective Function : MIN (Network Carrying Cost)

Constraints :Stores’ Service Level ≥

Target Service Level

Stores’ OUTL ≥

Minimum OUTL(Minimum OUTL = Minimum Presentation Quantity)

Variables :Echelons’ OUTLs

Objective Function, Constraints, and Variables


Finding Optimal Result

2-Year Random Daily Store Demand(Poisson)

.....

Store- Segment 1

Store- Segment 2

Store- Segment 3

Store- Segment12

RetailCo’sDC


OrdersOrders

Delivered Quantity

OrdersDelivered Quantity

Orders

2

Varying DC’s and Supplier’s OUTL

1

Varying Stores’ OUTLs


475 (45%)

245(23%)

345 (32%)

92 (34%) 51 (19%)

129 (47%)

34 (22%) 23 (15%) 95 (63%)

-

200

400

600

800

1,000

1,200

Aver

age

Inve

ntor

y (U

nit)

High Medium LowSales Unit Volume

StoresDCSupplier

Optimal Inventory DistributionOptimal inventory carrying cost can be achieved by having low inventory at the DC

Sales Unit VolumeHigh Medium Low

Supplier 100.0% 99.0% 91.2%DC 97.0% 86.8% 75.6%Stores 98.8% 99.0% 99.1%

Service Level Low service level at the DC


How an echelon’s inventory policy affects other echelons?

0

20

40

60

80

100

120

140

20 40 60 80 100 120 140 160 180DC's OUTL (Unit)

Ave

rage

Inve

ntor

y (U

nit)Supplier

Store

DC

65%

70%

75%

80%

85%

90%

95%

100%

20 40 60 80 100 120 140 160 180DC's OUTL (Unit)

Serv

ice

Leve

l

StoreSupplier

DC

Inventory decreases as the DC decreases OUTL

Inventory increases as the DC increases OUTL BUT only at the DC!!

DC’s service level plummets as well as Store’s service level as DC’s OUTL drops

ONLY DC’s service level increases as DC increases OUTL!!


-50

100150200250300350

95% 96% 97% 98% 99% 100%Store Target SLA

vera

ge In

vent

ory

(Uni

t)

SupplierDC

Stores

Network

80%

85%

90%

95%

100%

95% 96% 97% 98% 99% 100%Store Target SL

Serv

ice

Leve

l

DC

StoresSupplier

How store target SL impacts inventory distribution and echelons’ SLs ?

Stores and DC needs to have more inventory

Stores and DC needs higher service level BUT Supplier can relax its service level!!


92%

75%

60%

27%20%

11%

90%

73%

61%

47%

96% 95%99% 98%100%

87%97%99%

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

100%

0% 2% 4% 6% 8% 10% 12%

Probability of Service Disruption

Serv

ice

Leve

l

Supplier

DC

Stores

How supplier’s service disruption impacts echelons’ SLs?

Supplier’s service disruption has significant impact on the downstream echelon, especially at DC.1% Service disruption causes store’s SL to drop below target SL


ConclusionThe optimal carrying cost can be achieved by keeping low inventory and service level at the DC.This study shows INTERRELATIONs among the echelons in a multi-echelon network.This relationship should be considered when setting up the echelon inventory policy

Eliminate excess safety stockReduce service failure

Higher store target service level = Higher inventory at the downstream echelonService failure can have significant impact on the network service levels


Question?

Q u e s t i o nQ u e s t i o n

inventory optimization in a retail multi echelon …...average inventory (unit) supplier store dc...

Documents