Inventory Control

–
subject to known demand
Production and operation analysis
Nahmias S.
Types of Inventory

• Raw materials,
• Components; items which are not yet reached
completion in the production system such as
subassemblies,
• Work in process (work in progress); is inventory
waiting in the system to be processed or being
processed,
• Finished goods; final products which completed
all the processes.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 2

Motivation for Inventory

❑ It improves system performance by decoupling

parts of the system from one another,

❑ It allows production system to be built with

capacity less than the peak demand,

❑ It reduces the propagation of disturbances and

thus reduces instability and fragility of complex,
expensive systems,

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 3

Motivation for Inventory

❑ Economies of scale; it might be economical to

produce or to order in large amounts and store
for future use.
❑ Uncertainties; Demand changes, lead time
variability, supply chain uncertainties etc..
❑ Speculation; inventory may be held in
anticipation of a rise in their value or cost.
❑ Transportation; in transit inventories which are
applicable for pipeline industries.
❑ Smoothing; to resist against demand fluctuations

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 4

Some terminology

• Lead time refers to time between the arrival of a order

and the accomplishment of it. For example;
• Order lead time; time between the order given to the
supplier and the arrival of it to the company
• Production lead time; time from the start of a production
order to the completion.

• Review time; intervals of time for controlling of

inventories. E.g. Periodic review, Continuous review (in
supermarkets through barcode system)
• Excess demand; when demand cannot be met, they
are either backlogged(delivered later) or lost.
• Changing inventory; time might affect the utility. Short
shelf life, becoming obsolete etc…

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 5

Relevant costs
• Holding cost,
• Cost of providing the physical space to store items,
• Taxes and insurance,
• Breakage, spoilage, deterioration, obsolescence,
• Opportunity cost.

• Order cost, is related with the amount inventory

ordered and the number of occasions.

• Penalty cost refers to shortage cost or stockout

cost
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 6
ABC inventory classification method
Multiproduct systems

• Pareto effect in inventory management: a large

portion of the total value of sales is often accounted
for by a small number of inventory items.
• Typically, the top 20% of the items (class A) account
for about 80% of the annual sales value, the next 30%
of the items (class B) for the next 15% of sales and the
remaining 50% (class C) is for the last 5% of sales value.
• Class A- high annual dollar volume
• Class B-medium annual dollar volume
• Class C-low annual dollar volume

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 7

ABC analysis
• Tight management control of ordering
procedures is essential for Type A items.
• For Type B items, inventories can be
reviewed periodically
• Items can be ordered in small groups,
rather than individually.

• Type C items require the minimum

degree of control
• Parameters are reviewed twice
a year. Demand for Type C items
may be forecasted by simple
methods. The most inexpensive items
of type C can be ordered in large lot, to minimize number of
orders. An expensive type C items ordered only as they are
demanded.
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 8
Fundemantel questions of inventory
control
1. What items to stock?
Objectives of a business and the strategy to achieve
the objectives. e.g. range of stock offered by retailers
2. Where to stock the items?
Should all items be stocked everywhere or should
certain items be stocked in only a single location?
3. How much should be ordered when an order is
placed?
Many factors to be considered: demand rate, cost of
holding inventory, fixed cost of ordering, ..
4. When should an order be placed?

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 9

Economic Order Quantity (EOQ)

• The EOQ model is the simplest and most

fundamental of all inventory models.
• Simplicity and restrictive modelling assumptions
usually go together; the EOQ model is not an
exception.
• Environment: single-stage system with a single
item to stock that has a continuously constant
and known demand rate.
• But the model produces good results in many
situations and has been effectively employed in
automotive, pharmaceutical, and retail sectors
of the economy for many years.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 10

EOQ Model
Notations and assumptions underlying the model

• l - demand rate (units/year) –

• demand arrives continuously at a constant and known rate
• Shortages are not permitted(all demand is satisfied on time).

• c – unit order cost ($/unit)

• not counting setup or inventory cost

• K – fixed or setup cost to place an order ($)

• When order is placed, it arrives instantly. Order lead time is zero

• Q – Unknown size of the order or lot size

• h – holding cost ($/unit/unit time) (=i*c where i = interest rate)

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 11

 EOQ Model
1 Cycle
Inventory, I(t)

Slope = -l
Order arrival

2T 3T Time, t
T

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 12

EOQ Model

• Our goal is to minimize the average (averaged over

time) total cost/unit time by determining the ideal
order quantity, Q.

• The time between the placing of two successive

orders is the cycle or the reorder interval, T.
→ T= Q / l.

the lower the Q, the more orders we will have to place

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 13

 EOQ Model
Cost function
Cost function per unit time (usually a year) to be minimized:
• Purchasing cost for one year = c*l
• Number of orders placed (number of cycles) per year= l / Q
• Annual fixed cost of placing orders= K*(l /Q)
• Average inventory per cycle : (area of a triangle) / (cycle
length).
1
QT
2 Q Q
= =
T 2
T
• annual cost of holding inventory: h*Q/2

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 14

EOQ Model
Cost function
𝐾λ ℎ𝑄
G Q = + λ𝑐 +
𝑄 2

Annual Annual
fixed cost holding cost
Annual
purchasing
cost

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 15

EOQ Model
Cost function

𝐾λ ℎ𝑄
Min G Q = + λ𝑐 + where Q>0
𝑄 2
𝑑2 𝐺 2𝐾λ
G’’(Q)= 2 = > 0 𝑓𝑜𝑟 𝑄 > 0
𝑑𝑄 𝑄3

Since G’’(Q) >0, it follows that G(Q) is a

convex function of Q.
The optimal value of Q occurs where
G’(Q) =0.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 16

EOQ Model
Cost function

𝑑𝐺 𝐾λ ℎ
=− 2+ =0
𝑑𝑄 𝑄 2
Hence;

2𝐾λ
𝑄= optimal order quantity !
ℎ

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 17

EOQ Model
The average annual cost function, G(Q)

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 18

Lets say

• λ= 3000 units/year
• c=0,005 $/unit
• K = 0,001 $/order
• h = 6$/unit/year

2𝐾λ 2∗0,001∗3000
• 𝑄= = =1
ℎ 6

𝐾λ ℎ𝑄 0,001∗3000 6∗1
Cost G Q = + λ𝑐 + = + 3000 ∗ 0,005 + =21$
𝑄 2 1 2

Number of orders per year = λ/Q = 3000 orders. Model

proposes to have a seperate order for each unit (Q=λ )

WHY?

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 19

EOQ Model K: lower for the order is the goal
fixed value
Effect of order cost on optimal order quantity!

K=0,01
K=0,001
K=0,0001

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 20

EOQ Model
Relation between Q and λ

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 21

EOQ Model
Relation of Q and K

K
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 22
EOQ Model
Relation of Q and h

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 23

Nonzero order lead time
If order lead time constraint is relaxed, how can we deal with it?

Reorder point
Inventory, I(t)

for order lead

time of t
Slope = -l
Order arrival

R=?

T t 2T 3T Time, t

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 24

Nonzero order lead time
• New parameters;
R - Reorder point (units of item)
t - order lead time
• 𝑅 = l (units/unit time)*t (unit time)
• R represents the reorder point. When the inventory level
drops to R, order should be given for the next cycle.

• What if t > T ?
• Form the ratio t/T.
• Consider only the fractional remainder of the ratio, m:

R*= l*(m*T)

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 25

Economic Production Quantity (EPQ)

In EOQ model, it is assumed that orders arrive as a

complete lot from outside suppliers. What if we want
to produce them internally with a constant rate of
production and still benefit EOQ formula?

New parameter;

• P - production rate (units per year), where P> l

• All the other assumptions of EOQ are valid.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 26

Inventory, I(t)
EPQ Model
ECONOMIC PRODUCTION QUATITY
NOT ORDER AS BEFORE

Slope = P-l
Slope = -l
H

T1 T2 Time, t
T
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 27
EPQ Model
• Each order cycle is T: time between successive
production startups. T is comprised of two sub-
cycles: T=T1+T2, where T1 is production time or
uptime and T2 is downtime.
• The number of items produced (= consumed) each
cycle is the lot size, Q= l*T = P*T1.→ T1=Q/P.
• The maximum level of on-hand inventory is not Q
but H (H < Q). H/T1 = P-l → H=Q(1 - l/P).
• The average inventory per unit time: area of the
triangle/T. H

(H*T / 2) / T= H / 2.

9/2/2022 T
MG2029 Production Planning & Control Hakan Akillioglu 28
 EPQ Model
• The average annual cost function, Z(Q):
K H Kλ h Q
Z(Q) = cλ + + h = cλ + + (1 - λ P )
T 2 Q 2
• EPQ=Q* is found by dZ(Q)/dQ=0:
2Kλ
EPQ = Q =
*

h (1 - λ P )
2Kλ
• Defining h = h (1 - λ / P) EPQ = Q = *

h'

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 29

Quantity discount models
By now, unit cost of product is assumed to be constant,c,
being independent from the amount of purchase.
• Lets change it;
• Unit purchasing cost decreases with the order quantity Q.
• All the other assumptions of EOQ remain unchanged.

• Two types of quantity discounts are common:

• all units discounts and incremental quantity discounts.
• An example of all units discount:

Quantity purchased Per-unit price

0-100 $5
101-250 $4.50
251 and higher $4

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 30

Quantity discount models in EOQ
All unit discount

• Total purchasing cost for all units discount example:

What is the problem in this area?

Total purchasing cost

c2=4.50 c3=4
c1=5

q1=0 q2=100 q3=250

Order quantity, Q

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 31

Quantity discount models in EOQ
All units discount

• A more common discount contract.

• One or more price breakpoints defining changes in the
unit cost. Let m be the number of discount possibilities. Let
q1=0, q2, q3, .. , qm be the order quantities at which the
purchasing cost changes.
• Unit purchasing cost in the range [qj, qj+1) is cj.
• The average annual cost function:

Kλ Icj Q
Z j (Q) = +cj λ+ , q j  Q  q j +1
Q 2
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 32
Quantity discount models in EOQ
All units discount

• A family of cost functions indexed by j. The jth

cost function is defined for only those values of Q
in [qj, qj+1).

Z(Q1)

Total Z(Q2)
annual
cost Z(Q3)

q1=0 q2 q3
c1 c2 c3
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 33
Quantity discount models in EOQ
All units discount

Two observations follow:

• The average annual cost function is not continuous. It is

segmented such that each segment is defined over a
discount interval [qj, qj+1). → makes solving the problem
harder.

• The cost curve at the top corresponds to the highest

per-unit purchasing cost c1. The lowest curve
corresponds to the lowest per-unit purchasing cost cm.
The curves do not cross each other.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 34

Quantity discount models in EOQ
All units discount

Algorithm to determine Q*

• Step 1: Set j=m. Compute the EOQ for the mth cost
curve, denoted by Qm*:

• Step 2: Is Qm* ≥ qm ? If yes, Qm* is the optimal order

quantity and we are done. If not, the minimum cost
occurs at Q= qm for this segment, due to the convexity
and non-crossing properties of the cost functions.
Compute the cost corresponding to Q=qm.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 35

Quantity discount models in EOQ
All units discount

Let this cost be Zmin and Qmin = qm and go to Step 3.

• Step 3: Set j=j-1. Compute the EOQ for the jth

cost curve:

• Step 4: Is Qj* in [qj, qj+1) ? If yes, compute Z(Qj*)

and compare with Zmin. If Z(Qj*) < Zmin , Qj* is the
optimal order quantity; else Qmin is the optimal
order quantity. In either case we are done.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 36

Quantity discount models in EOQ
All units discount

Step 4 (cont’d)

Otherwise, if Qj* is not in [qj, qj+1), then the minimum

cost for the jth curve occurs at Q= qj, due to the
convexity and non-crossing properties of the cost
functions.

Compute the cost, Z(qj).

If Z(qj)< Zmin, then set Qmin=qj and Zmin=Z(qj).

If j≥2, go to Step 3; otherwise stop.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 37

Quantity discount models in EOQ
All units discount - Example

Example: An office supplies store sees a constant demand

rate of 10 boxes of pencil per week. Each box costs $5. If
the fixed cost of placing an order is $10 and the holding cost
rate, i, is 0.20 per year, determine the optimal order quantity
using the EOQ model. Assume 52 weeks per year.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 38

Quantity discount models in EOQ
All units discount - Example

• l = 10*52 = 520 units/year

• c = $5
• K = $10
• Yearly interest rate = 20%
• h = i*c = 0,2*5 = $1

2𝐾λ 2∗10∗520
• 𝑄= = = 102
ℎ 1

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 39

Quantity discount models in EOQ
All units discount - Example

The retailer gets an all units discount of 5% per box

of pencils if he purchases at least 110 boxes in a
single order. The deal becomes better if the
retailer purchases at least 150 boxes in which he
gets a 10% discount.

Should the retailer change the order quantity?

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 40

Quantity discount models in EOQ
• All units discount - Example
SOLUTION:
m=3. c1=5, c2=5(1-0.05)=4.75, c3= 5(1-0.1)=4.50.
q1=0, q2=110, q3=150.
j=3.
2 (520) (10)
Step 1: compute
Q3 = *
= 107.5
(0.20) (4.50)
Step 2: Since Q3* is not greater than or equal to q3=150, the
minimum cost occurs at Q=150. Its cost Zmin is:
(10)(520) (0.2)(4.50)(150)
Zmin = (4.50)(520) + + = 2442.17
150 2
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 41
Quantity discount models in EOQ
• All units discount - Example

Step 3: Set j=2. Compute

2 (520) (10)
Q2 =*
= 104.63
(0.20) (4.75)

Step 3: Q2* is not feasible; it is not in [110,150). The

minimum feasible cost occurs at Q= q2=110:

(10)(520) (0.2)(4.75)(110)
Z(q2 ) = (4.75)(520) + + = 2569.52
110 2

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 42

Quantity discount models in EOQ
All units discount - Example

This cost is higher than Zmin and so Zmin remains

unchanged.
Step 3: j=1. Q1*=101.98.
Step 4: Q1* is feasible; it is in [0,110). The minimum feasible
cost occurs at Q= q1=101.98
The cost is 2701.98 > Zmin.
→ The optimal solution is to order 150 units with a resulting
cost of $2442.17.
→ Observe that the algorithm stops as soon as a discount
is found for which Qj*is feasible.
→ So change the Q from 102 to 150 !
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 43
Quantity discount models in EOQ
Incremental quantity discount

1215
c3=$4
Total purchasing cost

c2=$4.50
c1=$5

q1=0 q2=100 q3=250

Order quantity, Q
Q=260
For Q=260 units, total purch. cost=(5)(100)+(4.5)(150)+(4)(10)=$1215
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 44
Quantity discount models in EOQ
Incremental quantity discount

• Differs from the all units discount contracts.

• As Q increases, the unit purchasing cost, cj, declines

incrementally on additional units purchased.

• Let m be the number of discount levels. Let q1=0, q2, q3, .. ,

qm be the order quantities at which the unit purchasing
cost changes.

• Unit purchasing cost in the range [qj, qj+1] is cj.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 45

Quantity discount models in EOQ
• Incremental quantity discount

• If Q units in the jth discount interval [qj , qj+1] are

ordered, the purchasing cost is:

C(Q) = c1 (q2 - q1 ) + c2 (q3 - q2 ) + ... + c j -1 (qj - q j -1 ) + c j (Q - q j ).

Rj , j≥2
C(Q) = Rj + c j (Q - q j ).
• The average unit purchasing cost for Q units is
C(Q)/Q.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 46

Quantity discount models in EOQ
• Incremental quantity discount

C(Q) R j qj
= + cj - cj .
Q Q Q
• The average annual cost function, Z(Q):

C(Q)
i Q
C(Q) λ
Z(Q) = λ+K + Q
Q Q 2
 Rj qj  λ i(R j + c j (Q - q j ))
=  + c j - c j λ + K +
Q Q Q 2

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 47

 Quantity discount models in EOQ
Incremental quantity discount

• Rearranging the terms in Z(Q) and naming it as Zj(Q):

λ ic jQ i(R j - c j q j )
Z j (Q) = c j λ + (R j - c j q j + K) + + for q j  Q  q j + 1
Q 2 2

• Then we have a family of curves, Zj(Q) for each j: each

curve valid for a given interval is convex and
differentiable. The curves Zj(Q) and Zj+1(Q) cross at qj+1.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 48

Quantity discount models in EOQ
Incremental quantity discount

Z1(Q1)

Z2(Q2)

Z3(Q3)

q2=500 q3=1000

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 49

Quantity discount models in EOQ
Incremental quantity discount

Algorithm to determine Q*
• Step 1: Compute the order quantity that
minimizes Zj(Q) for each j, which is denoted by
Qj* and obtained by setting d Zj(Q) /dQ=0

dZ j (Q) λ ic j
= -(Rj - c j q j + K) 2 +
dQ Q 2
2 (Rj - c j q j + K) λ
 Qj = *

i cj
9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 50
Quantity discount models in EOQ
Incremental quantity discount

• This step gives us a total of m possible order

quantities.

• Step 2: We check the feasibility of the potential

values for Q*, that is, qj ≤ Qj* ≤ qj+1 ? Disregard the
ones that do not satisfy this inequality.

• Step 3: Calculate the cost Zj(Qj*) for each

remaining Qj*. The order quantity Qj* that gives
the least cost is the optimal order quantity.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 51

Quantity discount models in EOQ
Incremental quantity discount - Example

Office supplies again. The retailer is offered an

incremental quantity discount.
c1=5, c2=4.75, c3=4.50.
q1=0, q2=109, q3=149.
Solution:
Step 1:
R1=0,
R2= c1(q2- q1)=(5)(109-0)=545,
R3= c1(q2- q1)+ c2(q3- q2)= R2 +(4.75)(149-109)=735.
Compute Qj* values.

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 52

Quantity discount models in EOQ
Incremental quantity discount - Example

Step 1: cont’d.

2 (R1 - c1q1 + K) λ 2 (0 - 0 + 10) (520)

Q1* = = = 101.98
i c1 (0.2)(5)

2 (R2 - c2q2 + K) λ 2 (545 - (4.75)(109) + 10) (520)

Q =
*
2 = = 201.94
i c2 (0.2)(4.75)

2 (R3 - c 3q 3 + K) λ 2 (735 - (4.5)(149) + 10) (520)

Q =
*
3 = = 293.41
i c3 (0.2)(4.5)

9/2/2022 MG2029 Production Planning & Control Hakan Akillioglu 53

 Quantity discount models in EOQ
• Step 2: We disregard Q2* since Q2* is not in [110,149].
Q1* and Q3* are feasible.
• Step 3: We compute the costs for Q1* and Q3*:
λ ic 1Q1* i(R1 - c 1 q1 )
Z(Q ) = c 1 λ + (R1 - c 1q1 + K) * +
*
1 +
Q1 2 2
520 (0.2)(5)(101.98) (0.2)(0 - 0)
= (5)(520) + (0 - 0 + 10) + + = 2701.98
101.98 2 2
λ ic 3Q*3 i(R3 - c 3 q 3 )
Z(Q ) = c 3 λ + (R3 - c 3q 3 + K) * +
*
3 + .
Q3 2 2
520 (0.2)(4.5)(293.41)
= (4.5)(520) + (735 - (4.5)(149) + 10) + +
293.41 2
(0.2)(735 - (4.5)(149))
= 2610.52
2

The optimal solution is Q* = 293.41 with a resulting cost of

$2610.52
9/2/2022 > $2442.17 (all-units
MG2029 Production discount
Planning case).
& Control Hakan Akillioglu 54