SCperf: An inventory management package for R

Marlene Marchena
[email protected]

Department of Electrical Engineering

Pontifical Catholic University of Rio de Janeiro - Brazil.
 Outline

ˆ Inventory
ˆ The basic Economic Order Quantity (EOQ) model
ˆ EOQ assumptions
ˆ Derivation of the model
ˆ Inventory models
ˆ What is SCperf?
ˆ EOQ() example
ˆ Bullwhip Effect (BE)
ˆ Measuring the BE
ˆ Measuring the BE for a generalized demand process
ˆ SCperf()
ˆ Why did we develop SCperf?
 Inventory

ˆ What is inventory?

Stock of items kept to meet future demand

 Inventory

ˆ What is inventory?

Stock of items kept to meet future demand

ˆ Why to hold inventory?

To protect himself against irregular supply and demand

 Inventory

ˆ What is inventory?

Stock of items kept to meet future demand

ˆ Why to hold inventory?

To protect himself against irregular supply and demand

ˆ Inventory Control Decisions

Objective: To minimize total inventory cost

Decisions:
ˆ How much to order?
ˆ When to order?
 EOQ assumptions

1. Instantaneous production,

2. immediate delivery,

3. deterministic demand,

4. constant demand,

5. known fixed setup costs,

6. no shortages are allowed,

7. single product.
 EOQ model
Notation:
D: demand per time unit, A: ordering or setup cost,
h: holding cost per unit and time unit, Q: batch quantity,
c: unit cost for producing or T : cycle time= Q/D
purchasing each unit.
 EOQ model
Notation:
D: demand per time unit, A: ordering or setup cost,
h: holding cost per unit and time unit, Q: batch quantity,
c: unit cost for producing or T : cycle time= Q/D
purchasing each unit.
hQ 2
Total cost per cycle = A + cQ +
2D
 EOQ model
Notation:
D: demand per time unit, A: ordering or setup cost,
h: holding cost per unit and time unit, Q: batch quantity,
c: unit cost for producing or T : cycle time= Q/D
purchasing each unit.
hQ 2
Total cost per cycle = A + cQ +
2D
A + cQ + hQ 2 /2D DA hQ
Total cost per unit time = = + cD + h
Q/D Q 2
 ∂TC D Q
We have that = A + h,
r ∂Q Q 2
2DA Qopt
then Qopt = and Topt =
h D

Reorder Point: order when the inventory position is equal to zero.

 Modification of the basic model

1. Instantaneous production,

2. immediate delivery,

3. deterministic demand,

4. constant demand,

5. known fixed setup costs,

6. no shortages are allowed,

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery,

3. deterministic demand,

4. constant demand,

5. known fixed setup costs,

6. no shortages are allowed,

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery, ⇐ Lags can be added

3. deterministic demand,

4. constant demand,

5. known fixed setup costs,

6. no shortages are allowed,

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery, ⇐ Lags can be added

3. deterministic demand, ⇐ Stochastic demand

4. constant demand,

5. known fixed setup costs,

6. no shortages are allowed,

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery, ⇐ Lags can be added

3. deterministic demand, ⇐ Stochastic demand

4. constant demand, ⇐ Time-varying demand

5. known fixed setup costs,

6. no shortages are allowed,

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery, ⇐ Lags can be added

3. deterministic demand, ⇐ Stochastic demand

4. constant demand, ⇐ Time-varying demand

5. known fixed setup costs, ⇐ Constraint approach

6. no shortages are allowed,

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery, ⇐ Lags can be added

3. deterministic demand, ⇐ Stochastic demand

4. constant demand, ⇐ Time-varying demand

5. known fixed setup costs, ⇐ Constraint approach

6. no shortages are allowed, ⇐ Shortages are allowed

7. single product.
 Modification of the basic model

1. Instantaneous production, ⇐ Finite production rate

2. immediate delivery, ⇐ Lags can be added

3. deterministic demand, ⇐ Stochastic demand

4. constant demand, ⇐ Time-varying demand

5. known fixed setup costs, ⇐ Constraint approach

6. no shortages are allowed, ⇐ Shortages are allowed

7. single product. ⇐ Multiple products

 What is SCperf?
An R package for inventory control.
 What is SCperf?
An R package for inventory control.
â Inventory models
 What is SCperf?
An R package for inventory control.
â Inventory models
o Economic Lot Size Models with Constant Demands
Economic Order Quantity, EOQ()
Economic Production Quantity, EPQ()
 What is SCperf?
An R package for inventory control.
â Inventory models
o Economic Lot Size Models with Constant Demands
Economic Order Quantity, EOQ()
Economic Production Quantity, EPQ()
o Economic Lot Size Models with Varying Demands,
Wagner-Whitin algorithm, WW()
 What is SCperf?
An R package for inventory control.
â Inventory models
o Economic Lot Size Models with Constant Demands
Economic Order Quantity, EOQ()
Economic Production Quantity, EPQ()
o Economic Lot Size Models with Varying Demands,
Wagner-Whitin algorithm, WW()
o Stochastics Inventory Models
Newsvendor model, Newsvendor()
 What is SCperf?
An R package for inventory control.
â Inventory models
o Economic Lot Size Models with Constant Demands
Economic Order Quantity, EOQ()
Economic Production Quantity, EPQ()
o Economic Lot Size Models with Varying Demands,
Wagner-Whitin algorithm, WW()
o Stochastics Inventory Models
Newsvendor model, Newsvendor()
â Safety Stocks, SS()
 What is SCperf?
An R package for inventory control.
â Inventory models
o Economic Lot Size Models with Constant Demands
Economic Order Quantity, EOQ()
Economic Production Quantity, EPQ()
o Economic Lot Size Models with Varying Demands,
Wagner-Whitin algorithm, WW()
o Stochastics Inventory Models
Newsvendor model, Newsvendor()
â Safety Stocks, SS()
â Inventory and Supply Chain Management (SCM)
The bullwhip effect, bullwhip() and SCperf()
 EOQ() function

Implements the basic (and with planned shortages) EOQ model

Example:
> EOQ(8000,12000,0.3)

Q T TVC
25298.22 3.16 7589.47
 The Bullwhip Effect (BE)
Definition: The BE is the increase of the demand variability as one
moves up the supply chain.
 Quantifying the BE
A common index used to measure the BE is:

Var (qt )
M=
Var (dt )
 Quantifying the BE
A common index used to measure the BE is:

Var (qt )
M=
Var (dt )

ˆ M = 1, there is no variance amplification.

 Quantifying the BE
A common index used to measure the BE is:

Var (qt )
M=
Var (dt )

ˆ M = 1, there is no variance amplification.

ˆ M > 1, the BE is present.

 Quantifying the BE
A common index used to measure the BE is:

Var (qt )
M=
Var (dt )

ˆ M = 1, there is no variance amplification.

ˆ M > 1, the BE is present.

ˆ M < 1, smoothing scenario.

 Quantifying the BE
A common index used to measure the BE is:

Var (qt )
M=
Var (dt )

ˆ M = 1, there is no variance amplification.

ˆ M > 1, the BE is present.

ˆ M < 1, smoothing scenario.

Zhang 2004:

PL PL
2 i=0 ψi ψj
M =1+ P∞j=i+1
2
j=0 ψj
 The model
Inventory model
ˆ Two stage supply chain
ˆ Single item with no fixed cost
ˆ OUT replenishment policy
ˆ MMSE as forecast method

Define:

dt : demand qt : order quantity

L: lead time α: theP desired SL
yt = D̂tL + z σ̂tL D̂tL = q Lτ =1 d̂t+τ
z: Φ−1 (α) σ̂tL = Var (DtL − D̂tL )
SSLT = z σ̂tL √
SS = zσd L

qt = yt − (yt−1 − dt ) = (D̂tL − D̂t−1

L ) + z(σ̂ L − σ̂ L ) + d
t t−1 t
 SCperf()

Computes the BE and other SC performance variables.

Usage: SCperf(ar, ma, L, SL)
Arguments:
ˆ ar : a vector of AR parameters,
ˆ ma: a vector of MA parameters,
ˆ L: is the LT plus the review period which is equal to one,
ˆ SL: service level, 0.95 by default.

Example:
> SCperf(0.95, 0.1, 2, 0.99)
bullwhip VarD VarLT SS SSLT z
1.5029 12.3077 5.2025 11.5419 5.3062 2.3264
 Why did we develop SCperf?

ˆ Educational purposes:
to offer to useRs, teachers, researchers and managers a free,
open-source, package for inventory control
 Why did we develop SCperf?

ˆ Educational purposes:
to offer to useRs, teachers, researchers and managers a free,
open-source, package for inventory control

ˆ Managerial purposes:
might be used as an alternative (or complement) to other
SCM commercial packages.
 Why did we develop SCperf?

ˆ Educational purposes:
to offer to useRs, teachers, researchers and managers a free,
open-source, package for inventory control

ˆ Managerial purposes:
might be used as an alternative (or complement) to other
SCM commercial packages.

ˆ The long-term goal of SCperf is to implement the last

research in inventory control theory as well as all the
state-of-the-art capabilities that are currently available in
commercial packages.
Thank you for your attention!

Marlene S. Marchena
[email protected]