You are on page 1of 96

MANAGEMENT OF INVENTORY SYSTEMS

Dynamic Inventory Problems under Certainty

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR

1
Sub Topics
1. General Characteristics, Q-system of Inventory Control, P-system of
Inventory Control, Determination of Economic Order Quantity (EOQ)
2. EOQ and Optimal Total Cost, Determination of Economic Production
Quantity (EPQ), Numerical Examples
3. EOQ with Planned Shortages, Numerical Examples, Determination of EOQ
with Quantity Discount
4. Determination of EOQ with Price Discount: Analytical and Graphical
Methods, Numerical Examples
5. Determination of Optimal Order Quantity under Constraints, Optimal
Policy Curve, Numerical Examples

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 2
Dynamic Inventory Problems under certainty

 General Characteristic
 Q-system of Inventory Control
 P-system of Inventory Control
 Determination of Economic Order Quantity (EOQ)

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 3
General Characteristics
• In this kind of inventory problems, more than one order is possible.

• Level of demand for the given item is known over the period of time
considered.

• Either the demand is expressed in total number of units or at some constant


rate over time. This rate may vary over time periods, but variable demand
rates are known with certainty.

• As demand is known with certainty, there is generally no need to consider


possibilities of overstock and understock.
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 4
General Characteristics
• Multiple orders are possible; we need to consider order cost to reflect
penalties with more than less orders.

• We need to consider a carrying or holding cost to reflect penalties resulting


from keeping a high level of stock over a period rather than a lower level.

• While we formulate the problem, we need to consider both ordering cost


and inventory carrying cost. These two costs are opposing to each other i.e.
if order cost is more, carrying cost is less, and vice-versa.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 5
General Characteristics
• For many materials used for construction, fabrication, or production
purpose, the demand is known with certainty. For example, inventories of
construction materials for a specific project need to be maintained and their
demand is known with certainty.

• For both Q-system and P-system of inventory control, their parameters can
be formulated with demand assumed to be known with certainty.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 6
General Characteristics
• The problem is also referred to as ‘Independent demand systems’ and as the
demand is known with certainty, the models prescribed to determine the
parameters of the inventory control system under consideration are
deterministic in nature.

• In order to determine an optimum inventory policy, data related to the


following three parameters are needed:
i. Demand levels or demand rates
ii. Relevant inventory-related costs (ordering and carrying cost)
iii. Lead times
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 7
General Characteristics
• Before we formulate the problems, the working and parameters of two
‘pure’ inventory control systems, P and Q, need to be known.

• Q-system of inventory control, also known as Fixed Order Size System (FOSS).

• P-system of inventory control, also known as Fixed Order Interval System


(FOIS).

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 8
Working of FOSS
• Flow diagram  Parameters: order quantity and
(Stock receipt)
Stock available reorder point
 Continuous review system
Demand occurs (units withdrawn)
 Applicable mainly for direct,
Determine stock position (on-hand + expensive and important items
on-order – backorders)
 Order quantity is fixed, but order
No Stock position ≤
interval is variable.
Reorder point
Yes
Issue replenishment order

9
Working of FOIS
• Flow diagram (Stock receipt)
Stock available
 Parameters: order interval and
Demand occurs maximum stock
Yes
 Periodic review system
Stock >
demand
No
 Applicable mainly for indirect,
No Review
period Backorders/Lost inexpensive and unimportant/less
arrived sales important items
Yes
Determine stock position (on-hand + on-order  Order interval is fixed, but order
– backorders)
quantity is variable.
Compute order quantity (maximum stock -
stock position)

Issue replenishment order

10
Q-system of Inventory Control: Determination of Order
Quantity
• This item has a very steady demand; the order quantity is replenished by the
supplier with a fixed lead time or instantaneously, and the same quantity is
demanded in each order cycle.

• The optimal order quantity to be determined is known as ‘Economic Order


Quantity’ (EOQ). This is also known as Wilson’s Lot Size Formula or Harris
Formula.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 11
Determination of EOQ

• For the given problem, the inventory profile is given below.


Q
Q : Order Quantity
Inventory B : Reorder Point (ROP)/
on hand Reorder Level (ROL)
B ab = cd = ef = lead time

.
a b
.c d
.e f time

Order Order
cycle cycle

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 12
• This inventory profile is known as ‘saw-tooth’ curve

• Let annual demand = S units


order quantity = Q
Number of orders per year = S/Q
order cost per order = Co
Inventory carrying cost = i, and
price per unit (unit price) = Cu

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 13
‘Cost versus Order Quantity’ Curve

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 14
i. Total order cost/year =
ii. Inventory carrying cost
At the beginning of cycle =
At the end of cycle =
Total inventory carrying cost/year =

iii. Total inventory cost/ year =

iv. Total cost = Fixed cost + variable cost


=
Hence, TC = total variable cost
=

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 15
• For minimization of TC,

i. Necessary condition :

ii. Sufficient condition:

• Total cost curve is a convex function.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 16
• We derive the EOQ formula under the following assumptions:

i. Lead time is constant or zero


ii. Demand is known and uniform
iii. Order quantity per order remains constant
iv. Unit price is constant
v. Order cost per order is constant
vi. Inventory carrying cost remains constant
vii. Single item
viii. Capital constraint is absent
ix. No limit on order quantity
x. No limit on number of orders

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 17
Dynamic Inventory Problems under Certainty

 EOQ and Optimal Total Cost


 Determination Economic Production Quantity (EPQ)
 Numerical Examples

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 18
EOQ and Optimal Total Cost
• EOQ can be expressed in different forms. It can be expressed in its
equivalent ‘economic order interval’. Its derivation is as follows:

• Let order period, t is in months, and N is the total number of orders per
year or number of order cycles per year.
• Hence, t = 12/N months N = 12/t

• Now, N = S/Q Q = S/N = St/12 (if time period is expressed in weeks and
we assume there are 52 weeks in a year, Q = S/N = St/12).
• Hence, TC = + = +

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 19
• Referring to the original EOQ formula with its total cost curve, we find that
the total cost is very insensitive with small change in the value of each of
the EOQ parameters.

• With , the optimal total variable cost, Eo is derived as

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 20
• Sensitivity of the TC with respect to the change in the value of one
parameter, say Co is determined as

• Similar form of equation you may derive for each of other parameters.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 21
• For getting the total cost expression, given Cu and S, we need to estimate
the value of Co and i. In many instances, there may be errors in this
estimation. It is imperative that the error should be as minimum as
possible , if not zero. The key question is: what is the impact of this error
on total cost?
• Let us analyze this question in the following manner:

Let actual values be i and Co , and corresponding considered values be k1i


and k2Co .

• Hence, Qo=

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 22
Actual total cost based on Qo =

1
¿ √ 2 𝑆 𝐶 𝑜 𝑖 𝐶𝑢
2 [√ √ ]
𝑘1
𝑘2
+
𝑘2
𝑘1

Actual optimal total cost = √ 2S 𝐶𝑜 𝑖𝐶𝑢

Difference, ∆ =
1
2 √ 2𝑆𝐶 𝑜 𝑖𝐶𝑢 [√ √ ]
𝑘1 𝑘2
+
𝑘2 𝑘1
− √2 𝑆 𝐶 𝑜 𝑖𝐶 𝑢

23
If 0.412 < k1/k2< 2.428 , < 10%

If 0.5 <k1/k2<2,
< 6%

• With this analysis, we may conclude about accuracy of the estimates, and if any
improved information system is needed for inventory management.

24
Determination of Economic Production Quantity
• The EOQ model as presented is for an inventory item that is purchased from
outside. A purchase or replenishment order is placed with order quantity, Q
stated.

• However for self-supply situation, EOQ model is not applicable.

• In a self-supply situation, the item is being produced internally rather than


procured from external supplier. When the production begins, a constant
number of units are supposed to be added to the inventory each day till the
time the production run is completed.
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 25
• During the production run, the item is purchased at a rate, referred to as
production rate. Simultaneously, the item is also demanded at a rate during
the production run time.

• Obviously, the production rate has to be higher than the demand rate, and
there is inventory build up.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 26
• Inventory profile – EPQ model

Q
Inventory Maximum
level inventory

p-d
B

tp LT Time
T

27
• For each production run for given production cost, the production system
(say, a machine) has to have a setup for which a setup cost is to be incurred.

• Let us assume,

Q = total (order) quantity


tp = production run time
p = production rate
d = demand rate
A = setup cost per setup
h = holding cost per unit per year

28
• Maximum inventory in a production cycle is given by

maximum inventory = tp (p - d)

tp = Q/p

Hence, average inventory =

• Total variable cost = setup cost + holding cost


=

29
• Setting the first derivative of the total cost expression with respect to the
decision variable, Q to zero, as a necessary condition, the optimal order
quantity, Q* is determined as

and the optimal total cost is given by

30
Numerical Example-1
A manufacturing company has determined from an analysis of its accounting
and production data for a certain part that (a) its demand is 9000 units per
annum and is uniformly distributed over the year, (b) its cost price Rs 2 per
unit, (c) its ordering cost is Rs 40 per order, (d) the inventory carrying charge
is 9 percent of the inventory value.
Further its known that lead time is uniform and equals 8 working days, and
that the total working days in a year are 300.
Determine:
(a) The economic order quantity, EOQ;
(b) The optimum number of orders per annum;

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 31
Numerical Example-1
(c) The total ordering and holding cost associated with the policy of ordering
an amount equal to EOQ;
(d) The re-order level;
(e) The number of days stock at re-order level;
(f) The length of inventory cycle;
(g) The amount of savings that would be possible by switching to the policy
of ordering EOQ determined in (a) from the present policy of ordering the
requirements of this part thrice a year; and
(h) The increase in total cost associated with ordering (i) 20% more, and
(ii) 40% less than the EOQ.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 32
Solution
We are given that D = 9000 units/year, A = Rs 40/Order, i = 0.09, c = Rs 2/ unit,
and, therefore, h = i × c = 0.09 × 2 = 0.18. Also, lead time = 8 working days, and
total working days in a year = 300.
(a) EOQ, Q* = = = 2000 units

(b) Optimum number of orders per year N* = D/Q*


= 9000/2000 = 4.5

(c) Total Variable cost, T(Q*) = =


= Rs 360

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 33
Solution
(d) Re-order level = lead time in days x demand per day
= 8 x 9000/300 = 240 units
(e) No. of days' stock at the re-order level = 8 (equal to lead time)
(f) Length of inventory cycle, T* = Q*/D = 2000/9000
= 0.222 year or 0.222 x 300 = 66.7 days
Alternatively, T* (in days) = Q*/demand per day
= 2000/30 = 66.7 days
(g) For the present policy of an order quantity = 3000 units,
Ordering cost = 40 x 3 = Rs 120
Holding cost = (3000/2) x 0.18 = Rs 270
T (3000) = 120 + 270 = Rs 390
Thus, saving in cost = Rs 390 - Rs 360 = Rs 30 per year.
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 34
Solution
(h) (i) Ordering 20% higher than EOQ:
Ordering quantity = (120/100) x 2000 = 2400 units.
With Q* = 2000 and Q = 2400, k = 2400/2000 = 1.2
We have,
T(Q)/ T(Q*) = 1/2 = 1/2 = 61/60
Thus, the cost would increase by 1/60th or 360 × 1/60 = Rs 6.
(ii) Ordering 40% lower than EOQ:
In such a situation, k = 1.40, and
T(Q)/ T(Q*) = 1/2 = 37/35
Thus the increase in cost would be 2/35th over the cost for EOQ, and
would equal 360 x 2/35 = Rs20.56.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 35
Numerical Example-2

A Contractor has to supply 10,000 paper cones per day to a textile unit. He
finds when he starts a production run, he can produce 25,000 paper cones
per day. The cost of holding a paper cone in stock for one year is 2 paise and
the setup cost of production run is Rs 18. How frequently should the
production run be made?

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 36
Solution
Assuming 300 working days in the year, we have D = 10,000 × 300 = 30,00,000
units, A = Rs18/set-up, h = 0.02/unit/year, p = 25,000 units/day, d = 10,000
units/day.
Accordingly, optimal set up quantity , Q* can be obtained as,
Q* =

= = 94,868 ≈ 95,000
Frequency of production runs can be fund as follows:
T* = Q*/d = 95,000/10,000 = 9.5 days
Thus, production run can be made after every 9.5 days.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 37
Dynamic Inventory Problems under Certainty

 EOQ with Planned Shortages


 Numerical Example
 Determination of EOQ with Quantity Discount

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 38
Determination of Economic Order Quantity with Planned
Shortages
• For certain items that are usually very expensive (high unit purchase price)
like fashion items (like jewellery, dress materials, etc), shortages of certain
amount (say, 10% or 20%) may be planned in order to reduce inventory
holding cost.

• Depending on the response of the customers, the shortage or stockout may


result in either backorders (customers not withdrawing orders) or lost sales
(customers withdrawing orders as the item’s availability is to be assured like
medicines, and other essential goods and products).

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 39
• Obviously, the shortages can be planned only for backordering case.

• Let us assume
Q = order quantity or lot size
S = shortage quantity planned
t1 = time when inventory on hand
t2 = time when there is a shortage
b = backordering cost per unit per year

• When the lot of Q units of the item is received, the customers whose orders
are pending would be supplied their requirements immediately and the
maximum inventory is (Q - S).
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 40
• Inventory profile with planned shortages is given below.

Q-S

d
Inventory

Q
level

𝑡2

t1 Time

S
T

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 41
• Holding cost during ordering cycle,

As (Q - S) = d and Q = Td, we have

or

Hence, holding cost during ordering cycle T

42
• Now, there are N number of order cycles per year, NT = 1 (year)
Annual holding cost =

• Shortage cost during ordering cycle,


as d = S/t2 = Q/T , t2 =
Hence, shortage cost per cycle =

43
• Annual shortage cost =

Total variable cost expression is given by

• Taking partial derivative of total variable cost expression with respect to Q


and S and setting them to zero as a necessary condition to satisfy, we get
the following expressions for optimal order quantity, Q* and optimal
shortage, S* :

44
Maximum interval level, M is given by

And total optimal variable cost, T(Q*) is given by

• If b>>h, order quantity, Q approaches original EOQ


• If h>>b, the backorder quantity, S is large

45
Numerical Example-1
A dealer supplies you the following information with regard to a product
dealt in by him:
Annual Demand : 10,000 units
Ordering cost : Rs 10 per order
Inventory carrying cost : 20% of the value of inventory per year
Price : Rs 20 per unit
The dealer is considering the possibility of allowing some backorder
(stockout) to occur. He has estimated that the annual cost of back-ordering
will be 25% of the value of inventory.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 46
Numerical Example-1
(a) What should be the optimum number of units of the product he should
buy in one lot?
(b) What quantity of the product should be allowed to be back ordered, if
any?
(c) What would be the maximum quantity of inventory at any time of the
year?
(d) Would you recommended to allow back-ordering? If so what would be
the annual cost saving by adopting the policy back-ordering?

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 47
Solution
According to given information, that D = 10,000 units/year, A = Rs 10/Order,
h = 20% of Rs 20 = Rs 4/unit/year and b = 25% of Rs20 = Rs 5/unit/year.
(a) Economic order quantity, (i) when stock-outs are not permitted,
Q* = = = 223.6 units
(ii) When back-ordering is permitted
Q* = = = 300 units
(b) Optimal quantity of the product to be back-ordered,
S* = = = 133 units

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 48
Solution
(c) Maximum inventory level,
M = Q* = 300 × = 167 units

(d) We have
TC (223.6) = = = Rs 894.43
TC (300) = = = Rs 666.67

Since the total cost with back ordering permitted is lower than when it is not
permitted, the dealer should accept the proposal for back ordering. This will
result in saving of 894.43 – 666.67 = Rs 227.76 per annum.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 49
Determination of Economic Order Quantity with Quantity
Discount
• In many situations, for a given item, a lower value of unit price is offered for
order amounts exceeding some given quantity.

• The classical EOQ model assumes that the unit price remains same
irrespective of the order size. However, in many instances, this may not be
the case, as a lower unit price may be quoted if the order is of a big size.

• In this case of price discounts, the order quantity is to be determined taking


into consideration the price levels for different quantity ranges.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 50
• There may be two situations – single price break or multiple price break.

• Examples :
1. Single price break:
Order Quantity (in units) Price per unit (Rs)
Q < 800 10.00
Q ≥ 800 9.00

In general, if Q < q1, C i = Co


Q ≥ q1, Ci = C 1

where, C1 < Co
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 51
2. Multiple price break:
Order quantity (in units) Price per unit (Rs)
Q < 300 10.00
300 ≤ Q <600 9.50
600 ≤ Q <1000 9.25
1000 ≤ Q <1500 9.00
Q ≥ 1500 8.75

In general; Co if Q < q1
C1 q 1 ≤ Q < q2
Ci = .
.
Cn-1 Q ≥ qn
Where, >>>………. >
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 52
• While determining the total cost in this case, as the unit price is a variable,
the purchase cost may vary against an order quantity, and hence, total cost
includes the purchase cost along with ordering cost and inventory holding
cost.

• We need to determine the order quantity at a particular price so that the


total cost is minimum.

• Let us first consider a single price break situation, and an analytical approach
to determine the order quantity.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 53
Dynamic Inventory Problems under Certainty

 Determination of EOQ with Price Discount: Analytical and


Graphical Methods
 Numerical Examples

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 54
Determination of EOQ with Price Discount
• Let Qo= EOQ without discount, and price discount starts at Q = k Qo , discount
rate = d, and hence, if initial unit price = Cu ,
the unit price with discount = (1-d) Cu .

• If total cost with discount is less than that neither discount, we opt for
discount; otherwise, we forgo discount.
• Hence,
if , the equality condition holds
when

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 55
• If less than kQo units get the discount, order the least amount necessary to
get it.

• If the amount which must be ordered to obtain the discount is greater than
kQo, discount should not be taken and Qo units should be ordered.

• Determination of order quantity can be explained with the graphical method.

• Let us assume, Cu1 = first price without discount


Cu2 = second price with discount
and Cu2 < Cu1
Qo = EOQ with Cu1
Qo2 = EOQ with Cu2 PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 56
• Four possible cases may be graphically illustrated.
Case-I:

As total cost against Qo2 (EOQ with


Cu2) is less than total cost against kQo
and Qo2 is feasible, decision is:
accept discount and order quantity
is Qo2

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 57
Case-II:
As total cost against kQo is
more than total cost against
Qo2, and Qo2 is feasible,
decision is: accept discount
and order quantity is Qo2

58
Case-III:
As total cost against kQo is less
than total cost against Qo , and
Qo2 is not feasible, decision is:
accept discount and order
quantity is kQo

59
Case IV:
As total cost against kQo is
more than total cost against
Qo, and Qo2 is not feasible,
decision is: do not accept
discount and order quantity is
Qo

60
• For all-units quantity discounts with multiple price breaks, the following
procedure is to be followed to determine the minimum total cost order
quantity for the given item:

Step-1 : Starting with the lowest unit price, calculate EOQ at each unit price,
until a valid EOQ is obtained.
Step-2: Calculate the total cost for the valid EOQ and for all other price break
quantities greater than the valid EOQ (a price-break quantity is the lowest
quantity for which the price discount is available).
Step-3: Select the quantity with the lowest total cost as the order quantity for
the given item.
• The procedure is presented with the following flow diagram:
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 61
Start

Calculate EOQ with lowest unit price

EOQ Yes Select EOQ as the order


valid quantity
No

Calculate EOQ with next higher unit price

EOQ No
valid
Yes
Calculate total cost for valid EOQ and all larger
price-break quantities

Select the order quantity with lowest total cost

End

62
Numerical Example-1
A diesel engine manufacturing company has planned its production
schedule for next year based on the forecasted demand, back orders and
plant capacity. Instead of manufacturing the piston, that goes in the final
product, the company has decided to buy the piston from ABC company. The
number of piston required are at the rate of 60 per day. Ordering costs have
been estimated at Rs 50 per order and the carrying cost fraction is 0.15. All
the assumptions of the basic EOQ model are applicable. The company
however can take advantage of one of several quantity discounts. The pricing
schedule of ABC is listed as follows:

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 63
Numerical Example-1

Quantity Ordered Unit Price (Rs)


0 - 1999 65
2000 - 4999 60
5000 – 10000 55
Over 10000 50

(a) What is the optimal order quantity?


(b) What is the minimum inventory cost?

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 64
Solution
Assuming 300 working days in the year, D = 300 × 60 = 1800 units, A = Rs 50
per order and I = 0.15. First we obtain EOQ using the most favorable unit price
of Rs 50,

Q* = = 490 units
It is feasible because the price of Rs 50 is available on an order of at least
10,000 units.
Next, we consider the price of Rs 55. for this,
Q* = = 467 units
This being also infeasible, we consider the price of Rs 60.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 65
Solution
Q* = = 447 units
This is also not feasible. Thus, we shall have to compute the EOQ value
considering the highest price of Rs 65. For this,
Q* = = 430 units
This is of course feasible. Now we shall determine the total inventory cost at
the feasible EOQ of 430 units, and at the various price-break quantities.

TC (430) = 18,000 × 65 + × 50 + × 0.15 × 65 = Rs 11,74,189

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 66
Solution
TC (2000) = 18,000 × 60 + × 50 + × 0.15 × 60 = Rs10,89,450

TC (5000) = 18,000 × 55 + × 50 + × 0.15 × 65 = Rs10,10,805

TC (10000) = 18,000 × 50 + × 50 + × 0.15 × 65 = Rs 9,37,590

The cost calculations shows that the total cost would be the minimum when an
order of 10,000 items is placed. Of course, the value 10,000 is taken only for
convenience. Thus, the optimal order size = 10,001 units.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 67
Numerical Example-2
A company uses annually 50,00 unit of an item each costing 1.20. Each order
costs Rs 45 and inventory carrying cost 15% of annual average inventory
value.
(a) Find EOQ.
(b) If the company operates 250 days a year, the procurement time in 10
days and safety stock is 500 units, find re-order level, maximum,
minimum and average inventory.
solution
(a) With A = Rs 45/order, D = 50,000 units, h = 15% of Rs 1.20 = 18, we have
EOQ, Q* = = 5,000 units

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 68
Solution
(b) Given annual demand, D = 50,000 units,
No of working days,N = 250,
Average daily demand = D/N = 50,000/250 = 200 units/day
Lead time = 10 days
Average DDLT = 200 10 = 2,000 units.

With safety stock, SS equal to 500 units, we have

ROL = SS + Average DDLT = 500 + 2,000 = 2,500 units


Maximum Inventory = SS + EOQ = 500 + 5,000 = 5,500 units
Minimum Inventory = SS = 500 units
Average Inventory = SS + 1/2 EOQ = 500 + 2500 = 3,000 units.
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 69
Numerical Example-3
The annual demand for a product is 3,600 units, with an average of 12 units
per day. The lead time is 10 days. The ordering cost per order is Rs 20 and
the annual carrying cost is 25% of the value of the inventory. The price of the
product per unit is Rs 3.
(a) What will be the EOQ?
(b) Find the purchase cycle time.
(c) Fine the total inventory cost per year.
(d) If a safety stock of 100 units is considered necessary, what will be the re-
order level and the total annual cost of inventory which will be relevant
to inventory decision.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 70
Solution
Given D = 3600 units, A = Rs 20/Order, i = 0.25, c = Rs 3/ unit, and lead time =
10 days. since the demand is uniform at 12 units per day, the total number of
working days in the year = 3600/12 = 300. We have,
(a) EOQ, Q* = = = 438 units approx.
(b) Purchasing cycle time, T* = Q*/Demand per day
= 438/12 = 36.5 working days
(c) Total inventory cost per year = 3600 × 3 + (3600/438) × 20 + (438/2) ×
0.75
= Rs 11,128.6
(d) Re-order level = Safety stock + Lead time demand
= 100 + 12 × 10 = 220 units

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 71
Solution
Total coat relevant to inventory decisions:
Holding cost = Average Inventory Unit Holding cost
Average Inventory = safety stock + 1/2 EOQ = 100 + 1/2 438 = 319 units

Total Holding cost = 319 0.75 = Rs 239.25

Total Ordering cost = No of Orders Ordering cost


= (3600/438) 20 = 164.38

Total Cost = 239.25 + 164.38 = Rs 403.63

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 72
Dynamic Inventory Problems under Certainty

 Determination of Optimal Order Quantity under Constraints

 Optimal Policy Curve

 Numerical Examples

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 73
Optimal ordering policy for more than one item
• In the problem formulations as mentioned in previous lecture sessions, two
specific assumptions we have made:
(i) Single item: formulae are applicable for one item.
(ii) No constraints: you refer to the assumptions related to classical EOQ
formula.

• In majority of the cases, we come across a situation where more than one
item (multiple items) may be considered jointly while determining optimal
order quantity itemwise, and there may be one or more of the constraints of
varieties under which the ordering policy is to be developed.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 74
• These constraints are of many types:
(i) restriction on the number of orders,
(ii) restriction on inventory investment,
(iii) restriction on storing space for inventory, etc.
• The question is: how to formulate inventory problems under constraints?
• Let us discuss the problem citing an example.
• A work unit of a plant maintains inventories of five items:
Item Si Cui
1 600 3
Si = annual demand for item i
2 900 10
Cui = unit price for item i
3 2400 5
4 12000 5
5 18000 1

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 75
• If for each item, the current ordering policy is to order each item once per
month, the number of orders per year and average inventory (assuming
uniform demand rate) for each item can be calculated as
Item Orders per year Average inventory(Rs)
1 12 75
2 12 375
3 12 500
4 12 2500
5 12 750
Total 60 4200

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 76
• For example, for item 2, average inventory = = 375

• Suppose, Co = 10 and i = 0.12 (same for all the given items). We calculate the
total variable cost (ordering and inventory carrying cost) for the current
inventory policy as
TVCexisting = 60 × 10 + 0.12 × 4200 = Rs 1104
• Suppose, we use the classical EOQ for the given items. The optimal order
size and orders per year for each item can be calculated using the following
expression of EOQ:
= × =
= EOQ in monetary terms

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 77
Item Optimal order size (in Rs) Orders per year
1 548 3.28
2 1225 7.35
3 1414 8.49
4 3162 18.98
5 1732 10.39
Total 8081 48.49

•Here, the orders per year for an item is calculated as

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 78
• Optimal total variable cost for this EOQ-based ordering policy is calculated as
TVCoptimal = 10 × 48.49 + 0.12 × 8081/2 = Rs 970
• Hence, there is a decrease of (1104 - 970) = Rs 134, around 12% in the total
variable cost with the use of EOQ-based ordering policy.
• Let us now go for further analysis of this problem. Certain important
observations we have out of the results obtained: average inventory
investment is Rs 4040 and the number of orders is 48.49. Now, in many
situations, there may be shortage of working capital and hence, an
investment of Rs 4040 in inventory may not be possible. There may also not
be possible to place 48.49 orders (ignoring the fractional value of order
number) by the existing purchase department because of manpower
shortage or non availability of adequate infrastructure.
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 79
• How to consider these constraints in the formulation of the problem so that
we are able to determine optimal ordering policy of an inventory item?

• The problem under constraint can be classified under two categories:

i. Minimize total cost subject to meeting one restriction/constraint


ii. Minimize total cost subject to meeting more than one
restriction/constraint simultaneously

• You need to determine the values of the decision variables.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 80
• Let us first consider the first type of problem.

• With the given dataset, let there be a restriction on average inventory


investment to Rs3,000. How to determine optimal order quantity for each item?

• We need to use the Lagrangian method to formulate and solve the problem
(minimize the total cost subject to restriction on inventory investment).

• The formulation is as follows:


Minimize TC = 10 + 0.12
such that = 3000

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 81
• We form Lagrangian, L as

L = 10 + 0.12 + λ

where λ = Lagrangian multiplier

• For minimization of L over Qi and λ, we take partial derivatives of L with


respect to Qi and λ and set them to zero:
= - 10 + 0.06 + λ = 0
= =0

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 82
• Solving these two simultaneous equations, we get

= and = 20 - 0.12

• For determining λ, we compute from the given dataset.

• For the given dataset, we get = 625.95


• With this value used in the expression of λ, we get λ = 0.09767 and the
generalized expression of Qi,Rs (order size in monetary units) is given by
= = 9.585

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 83
• Hence, itemwise order size and number of orders per year are computed as
Item Order size (in Rs) No of orders per year
1 406.70 4.43
2 909.30 9.90
3 1050.00 11.43
4 2348.00 25.56
5 1286.00 14.00
Total 6000 65.32

and hence, total variable cost = 10 × 65.32 + 0.12 × 6000/2 = Rs 1013, an


increase of Rs 43 in comparison with EOQ-based ordering policy (Rs 970).

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 84
• One important issue you need to consider at this stage: How would you
formulate the problem when estimates of ordering and inventory carrying
costs are not reliable? The problem can be formulated in two ways. The
procedure to be followed is as follows:

Total average inventory = TI =

Total number of orders = TO =

• In the first case, there may be restriction on the number of orders. For
example, the existing inventory policy for five items considered, the total
number of orders is 60.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 85
• We formulated the problem as
Min TI =
s.t. TO = = 60
• Hence, Lagrangian, L = + λ
= - =0
= =0
= = 54.42
= = = 10.43

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 86
• Itemwise, the order size and number of orders per year are computed as
Item Order size (in Rs) Orders per year
1 442.70 4.07
2 989.80 9.09
3 1142.80 10.50
4 2554.90 23.48
5 1399.30 12.86
Total 6529.50 60.00

• Average inventory is 6529.50/2 = Rs 3265 : reduced by Rs 935 or 22.3% reduction in


comparison with the existing inventory policy.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 87
• In the second case, there may be restriction on the average inventory
investment while minimizing the number of orders. For example, the existing
inventory policy for the five items under consideration, the total average
inventory investment is restricted to Rs 4200.

• We formulate the problem as

Minimize TO =

s.t.TI = = 4200

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 88
• Hence, Lagrangian, L = + λ
=- +=0
==0
=
λ=
• For the given dataset, λ = 0.0111 and Qi,Rs = 13.42

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 89
• Itemwise the order size and number of orders per year are computed as
Item Order size (in Rs) Orders per year
1 569.40 3.16
2 1273.10 7.07
3 1469.90 8.16
4 3286.40 18.26
5 1800.00 10.00
Total 8398.80 46.65

• Average inventory remains at 8398.80/2 = 4200 (with rounding errors) with number
of orders reduced to 46.65, 22.3% reduction in comparison with the number of
orders in the existing ordering policy.

90
• Hence, different solution are obtained while we deal with problems under
constraints. However, there is a general relationships between TI and TO. Let
us work out this relationship. For i number of items with j as inventory
carrying cost considered,
• We have TI = =
=
TO = =
• Hence, TI × TO = =K, a constant
and TI/TO =

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 91
• As has been found, TI × TO = K is an equation for rectangular hyperbola. This
equation represents the optimal policy curve :

Total
Inventory Z2
(TI)
B b
X
Z1
Y a
C

Total orders (TO)

92
• Any point on the curve (like A, B, or C) represents optimal policy with
corresponding values of TI and TO known (an inverse relationship between
them).
• How to use optimal policy curve for improving inventory control system?
• Any inventory control system may be evaluated with the two parameters: TI and
TO. Suppose, existing inventory policy is represented by point x or point y. They
are not optimal policy as x or y does not fall on the optimal policy curve.
• In order to reach to the optimal policy curve, you have the following three
alternatives:
(i) Change TO from x to Z1 , keeping TI constant
(ii) Change TI from y to Z2, keeping TO constant
(iii)Change both TI and TO simultaneously, for example, follow the path xa or
yb as shown in Figure.
PROF PRADIP KUMAR RAY
DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 93
• Among the three alternatives, alternative (i) and (ii) is preferred to alternative
(iii) as controlling one of the two parameters is always easier than controlling
both the parameters simultaneously.

• In case, the two or more constraints need to be considered simultaneously,


we form Lagrangian consisting minimization function in terms of decision
variable (order quantity) and constraints-wise multipliers (λ1, λ2, ......, λmwith m
number of constraints). The optimal ordering policy is determined by
considering partial derivatives of L with respect to order quantity (Qi)and each
of the multipliers considered. ( = 0, = 0, = 0, ………, = 0) and solving the
corresponding simultaneous equations.

PROF PRADIP KUMAR RAY


DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING
IIT KHARAGPUR 94
List of Reference Textbooks
• Starr, M K and Miller, D W, Inventory Control: Theory and Practice,
Prentice Hall.

• Tersine, R J, Principles of Inventory and Materials Management, PTR


Prentice Hall.

• Silver, E A, Pyke, D F and Peterson, R, Inventory Management and


Production Planning and Scheduling, John Wiley.

• Vohra, N D, Quantitative Techniques in Management, Tata McGraw Hill

Prof Pradip Kumar Ray


Department of Industrial and Systems Engineering 95
Thank You!!

96

You might also like