Supply Chain Analytics

Dr. Gopal Kumar

gopalkumar@iimraipur.ac.in

IIM Raipur
10 Oct 2024

Gopal Kumar (IIM Raipur) Supply Chain Analytics 10 Oct 2024 1/34
Retail analytics (Ordering policy)

Simulation types

❑ Discrete event simulation (DES)

✓ Customers arriving and served by a restaurant
✓ Production of cars
✓ Supply chain of an items
✓ Car service station
✓ Examining patients in a hospital

❑ Continuous simulation

Gopal Kumar (IIM Raipur) Supply Chain Analytics 14 Oct 2024 2/14
Choosing Distribution for Uncertain Parameters
By carefully selecting the appropriate probability distribution,
you can create more accurate models
Distribution Use Cases When to Use
Lead times, demand,
Normal Symmetrical variation around a mean
production times
Lead times, project Limited data but known min, max,
Triangular
timelines and most likely
Equal likelihood of outcomes across a
Uniform Demand, lead times
range
Time between failures,
Exponential Time between independent events
demand intervals
Count of events within a fixed time
Poisson Order arrivals, stockouts
period
Demand spikes, supplier Skewed distributions with rare but
Lognormal
lead times large peaks

Supplier reliability, Modeling probabilities within

Beta
failure probabilities bounded range (0 to 1)

Gopal Kumar (IIM Raipur) Supply Chain Analytics 10 Oct 2024 3/34
Choosing Distribution for Uncertain Parameters
By carefully selecting the appropriate probability distribution,
you can create more accurate models

Log-
Distribution AD p Value LRT AIC
Likelihood
Weibull -190.3 0.237 >0.25 384.6
Weibull –
Three -189.4 0.355 >0.25 0.174 384.7
Parameter
LogNormal -196.2 1.385 0.001 396.3
LogNormal –
Three -192.9 0.514 0.189 0.011 391.9
Parameter

High value represents a better fit

Likelihood ratio test (LRT): Lower value represents an improvement in the fit
(with additional parameter)
Akaike information criterion (AIC): Smallest AIC value is usually the preferred model
Gopal Kumar (IIM Raipur) Supply Chain Analytics 10 Oct 2024 4/34
Network Design Decisions

Network Design Decisions

Gopal Kumar (IIM Raipur) SCA 10 Oct 2024 5/34

Network Design Decisions
Supply Chain Stage and Distribution Network
❖ For different products and in different geographic area,
product distribution accounts for 10-35% of the revenue.

❖ It plays a significant role in satisfying customers.

Gopal Kumar (IIM Raipur) SCA 10 Oct 2024 6/34

Network Design Decisions
Supply Chain Stage

All these stages need not to be present

Gopal Kumar (IIM Raipur) SCA 10 Oct 2024 7/34
Network Design Decisions
• Facility role
– What role, what processes?

• Facility location
– Where should facilities be located?

• Capacity allocation
– How much capacity at each facility?

• Market and supply allocation

– What markets? Which supply sources?

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 8/34

Factors Influencing Network Design Decisions
• Strategic factors
• Technological factors
• Macroeconomic factors
– Tariffs and tax incentives
– Exchange-rate and demand risk
– Freight and fuel costs
• Political
• Infrastructure factors
• Competitive factors

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 9/34

Models for Facility Location and Capacity Allocation
Important information
– Location of supply sources and markets
– Location of potential facility sites
– Demand forecast by market
– Facility, labor, and material costs by site
– Transportation costs between each pair of sites
– Inventory costs by site and as a function of quantity
– Sale price of product in different regions
– Taxes and tariffs
– Desired response time and other service factors

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 10/34

Logistic Cost

Total logistic cost =

𝐼𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦 𝑐𝑜𝑠𝑡
+
Facility cost
+
Transportation cost.

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 11/34

Transportation Cost and Number of Facilities

EoS

EoS
Less
dis
EoS tan
ce

EoS

EoS Less
dis
tan
EoS ce

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 12/34

Costs and Number of Facilities

Inventory

Facility costs
Costs

Transportation

Number of facilities

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 13/34

Cost Buildup as a Function of Facilities

Total Costs
Cost of Operations

Percent Service
Level Within
Promised Time
Facilities
Inventory
Transportation
Labor

Number of Facilities

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 14/34

Framework for Network Design Decisions
.

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 15/34

Network Optimization Models
• Allocating demand to production facilities
• Locating facilities and allocating capacity

Key Costs:

• Fixed facility cost

• Transportation cost
• Production cost
• Inventory cost
• Coordination cost

Which plants to establish? How to configure the network?

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 16/34

Network Optimization Models
Case 1 (Exercise)
.
Plant Location with Multiple Sourcing
Sun Oil
Inputs - Costs, Capacities, Demands
Demand Region
Production and Transportation Cost per 1,000,000 Units Fixed Low Fixed High
Plant N. America S. America Europe Asia Africa Cost ($) Capacity Cost ($) Capacity

N. America 81 92 101 130 115 6,000 10 9,000 20

S. America 117 77 108 98 100 4,500 10 6,750 20

Europe 102 105 95 119 111 6,500 10 9,750 20

Asia 115 125 90 59 74 4,100 10 6,150 20

Africa 142 100 103 105 71 4,000 10 6,000 20

Demand 12 8 14 16 7
Aggregate level data
Gopal Kumar (IIM Raipur) Network Design 14 Oct 2024 17/15
Network Optimization Models
Case 1 (Exercise)
.
Plant Location with Multiple Sourcing
Sun Oil

Gopal Kumar (IIM Raipur) Network Design 14 Oct 2024 18/15

Capacitated Plant Location Model
Casen 1 (Exercise)
= number of potential plant locations/capacity
m. = number of markets or demand points yi = 1 if plant i is open, 0 otherwise
D j = annual demand from market j xij = quantity shipped from plant
K i = potential capacity of plant i i to market j
f i = annualized fixed cost of keeping plant i open
cij = cost of producing and shipping one unit from plant i to market j (cost
includes production, inventory, transportation, and tariffs)
n n m
Minå f i yi + å åc x ij ij
i=1 i=1 j=1

subject to
n

åx ij
= D j for j = 1,...,m
i=1
m

åx ij
= K i yi for i = 1,...,n
j=1

yi Î {0,1} for i = 1,...,n, x ij ³ 0

Gopal Kumar (IIM Raipur) Network Design 14 Oct 2024 19/15
Capacitated Plant Location Model
Case 1 (Exercise)
.

Gopal Kumar (IIM Raipur) Network Design 14 Oct 2024 20/15

Capacitated Plant Location Model
Case 1 (Exercise)
.

Gopal Kumar (IIM Raipur) Network Design 14 Oct 2024 21/15

Capacitated Plant Location Model
Case 1 (Exercise)
.

Gopal Kumar (IIM Raipur) Network Design 14 Oct 2024 22/15

Network Optimization Models
Case 2 (Exercise)
A manufacturing firm is designing supply chain network and
has selected three potential sites, Delhi, Mumbai, and Chennai
to serve markets of Kolkata, Bengaluru, Hyderabad, and
Nagpur. The firm estimates cost of building plant, production,
transportation, and tax rate for each manufacturing location
which are shown below:

A plan under discussion is if they build high (700 unit) or low

(400 unit) capacity plant, will they be able to meet the demand
in a cost-effective fashion? To build low and high-capacity
plant, the firm estimates a cost of ₹70,000 and ₹95,000
respectively
Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 23/34
Network Optimization Models
Case 3

(a) In case 1, if demand of both Asia and Africa increases

by 20% each year and it becomes stable after five
years, what should be a good supply chain design plan
considering 20% discount rate or opportunity cost?

(b) What if in case 1, demand of all market increases by

20% each year?

Gopal Kumar (IIM Raipur) Network Design 10 Oct 2024 24/34

 .

Thanks!

Gopal Kumar (IIM Raipur) Supply Chain Analytics 10 Oct 2024 25/34