2012

Inventory Models
Probabilistic

NATIONAL UNIVERSITY 'SANTIAGO

ANTÚNEZ DE MAYOLO
ACADEMIC PROFESSIONAL SCHOOL OF
ADMINISTRATION

ESPINOZA MENA CINTYA

ROJAS LEÓN MELISSA
Tello Valladares Aracely
TORRES LUCIANO JOSÉ
28/05/2012
 Probabilistic Inventory Models

INDEX
INTRODUCTION..............................................................3

OBJECTIVES

CONCEPT OF INVENTORY

ADVANTAGES OF AN INVENTORY SYSTEM....................................................5

INVENTORY CONTROL...................................................................................6

1 PROBABILISTIC INVENTORY MODELS...............................................8

1.1 Continuous review model.................................................................................... 8

1.1.1. Probabilistic Model of Economic Order Quantity.............................. 9
1.1.2. Probabilistic Model of Economic Order Quantity............................ 12
1.2 Single period model..................................................................................... 16
1.1.1. Unprepared Model ............................................................................... 17
1.1.2. Model with Preparation
1.3 Multiple Period Model..................................................................................... 21

EXERCISES................................................................................................................25

2
 Probabilistic Inventory Models

INTRODUCTION

Inventory, like accounts receivable, represents a significant proportion of the

assets in most companies that require substantial investments. By

Hello, the administrative practices that result in minimizing the percentage of

Total inventory can represent significant savings in money.

3
 Probabilistic Inventory Models

OBJECTIVES

The objective of inventory models is to present some methods that help to

achieve good management of inventories and an efficient relationship with them

Financial Administration.

CONCEPT OF INVENTORY

Inventories are a bridge between production and sales in a company.

The manufacturer balances the production line if some machines operate.

at different volumes of others, as one way to compensate for this imbalance is

providing temporary inventories or banks.

The inventories of raw materials, semi-finished products, and finished products

absorb the slack when sales or production volumes fluctuate, which

gives us another reason for inventory control.

These tend to provide a steady flow of production, facilitating their

programming.

Raw material inventories provide flexibility to the company's purchasing process.

Without them, the company is in a 'hand to mouth' situation.

4
 Probabilistic Inventory Models

raw materials strictly necessary to maintain the production plan, that is,

buying and consuming.

ADVANTAGES OF AN INVENTORY SYSTEM

With him, the company can carry out its production and purchasing tasks while saving money.

resources, and also serve their customers more quickly, optimizing all the

company activities. However, there is a disadvantage: the cost of

maintenance; since the cost of capital, the cost of storage, must be considered,

opportunity cost caused by absence, and others.

Both inventory and accounts receivable must be increased as far as the

the savings result is greater than the total cost of maintaining additional inventory.

The efficiency of the inventory system process is the result of good

coordination between the different areas of the company, having as premises its

general objectives.

INVENTORY CONTROL

The efficiency of inventory control can affect the operational flexibility of the

company.

Two essentially identical companies, with the same amount of inventory, but with

large differences in the degrees of flexibility of their operations, can have

5
 Probabilistic Inventory Models

imbalanced inventories, mainly due to inefficient controls of these. This

it causes them to encounter an abundance of some material at a certain moment

and lack another.

Finally, these deficiencies have negative effects on utility. In other words,

the ineffectiveness of inventory control for a given level of flexibility affects the

amount of investments required, that is, to lower efficiency in the system of

inventory control, greater investment needs. Consequently, the high

Investments in inventories will have an adverse impact on the company's profit.

Given the importance of an inventory control system, it is worth mentioning these.

generalObjectives

Minimize investment in inventory.

Minimize storage costs.

Minimize losses due to damage, obsolescence, or perishable items.

Maintain sufficient inventory so that production does not lack raw materials.

primes, parts, and supplies.

Maintaining an efficient transportation of inventories, including the functions of

dispatch and receipt.

Maintain an efficient inventory information system.

Provide reports on the inventory value to accounting.

Make purchases in a way that allows for economical acquisitions and

efficient.

Make forecasts about future inventory needs.

6
 Probabilistic Inventory Models

It is not possible to achieve all these objectives; certain steps must be taken to achieve them.

concessions. There are several conditions that hinder the achievement of these objectives. Rather

that represent problems that can be solved, these conditions are always present

presents and tends to frustrate effective inventory control.

The constant change in the supply-demand relationship frustrates effective control of the

inventory.

1. PROBABILISTIC INVENTORY MODEL

This topic refers to stochastic inventory models, in which the

demand is described by a probability distribution. The models that

they are classified, in the broad sense, into situations of continuous review and

periodic. The periodic review models include both single-period cases.

like from various periods. The proposed solutions range from the use of a version

Probabilistic determination of the economic order quantity up to more cases

complexes that are solved with dynamic programming. The probabilistic nature of

demand leads to complex models that may not be useful in practice. Without

embargo on publications good inventory implementations have been reported

probabilistic.

7
 Probabilistic Inventory Models

1.1 CONTINUOUS REVIEW MODELS

There are two models, the first is a 'probabilized' version of the deterministic EOQ.

that uses stabilizing inventories to explain probabilistic demand, the second

a more accurate probabilistic EOQ, which includes probabilistic demand in a way

directly in the formulation.

1.1.1 "PROBABILISTIC" MODELS OF ECONOMIC QUANTITY

ORDER

The size of the stabilizing stocks is determined such that the probability

of stock exhaustion during the delivery time (the period between placing

and receiving an order) does not exceed a predetermined value.

Figure 1.1

Inventory reserve tax on the classic CEP model

8
 Probabilistic Inventory Models

Sea:

LDelivery time between placing and receiving an order.

XL Random variable that represents demand during the delivery time.

µL Average demand during the delivery time.

σL Standard deviation of demand during lead time.

BSize of the reserve existence.

α Standard deviation of demand during lead time.

The main hypothesis of the model is that XL, the demand during the lead time L,

it has a normal distribution, with mean µL and standard deviationσL , that is, N (µL,σL ).

Figure 1.1 shows the relationship between the inventory reserve B and the parameters of

deterministic model of CEP, which includes the delivery time L, the average demand

during the time µL and the CEP, economic order quantity and*. Note that L must

be equal to the actual delivery time.

The formulation of the probability used to determine B can be written as:

P{XL≥ B+ µL≤ α}

XL can be converted into a standard normal random variable N(0,1) with the following

substitution

X L− µ L
z=
σL

9
 Probabilistic Inventory Models

So:

B
P{z ≥ }≤ α
σL

Figure 1.2

Probability that the stock will run out,P z ≥ Kα{= α }

Figure 1.2 defines aK α that is determined with the standard normal table, in such a way

what:

P{z ≥ Kα = α}

Consequently, the size of the reserve must satisfy

B ≥σL K α

Demand during the delivery timeLis commonly described with a function of

probability density per unit of time, that is, per day or week; based on

the one that allows determining the distribution of demand duringL. Given that the

10
 Probabilistic Inventory Models

demand per unit of normal time, with meanDand standard deviationσ, the average

µLand the standard deviationσL from the demand, during the delivery timeL , you

calculated as:

µL= DL

σL= √σ L2

In the formula ofσLit is required thatLlet it be an integer value.

Exercise:

Regarding the determination of the neon light inventory policy, it was determined that the

economic order quantity of CEP = 100 units. If the daily demand is normal,

with an average D = 100 lights and a standard deviation σ = 10 lights, that is, N (100, 10),

determine the size of the reserve so that the probability of running out of stock

less than = 0.05 and the effective delay time is L = 2 days.

µL= DL = 100 ×2= 200 units

σL= √σ L2 = 10√ ×214.14

2 units

B≥ = 14.14 × 1.645 ≈ 23 neon lights

The optimal inventory policy with a reserve B establishes purchasing 1000 units.

whenever the inventory level drops to 223 units.

11
 Probabilistic Inventory Models

1.1.2 PROBABILISTIC MODEL OF ECONOMIC ORDER QUANTITY

This model allows for shortfalls in demand, the policy requires ordering the quantity and

whenever the inventory falls to level R. As in the deterministic case, the level of

reorder R is a function of the delivery time, between placing and receiving an order. The

optimal values of 'y' and 'R' are determined by minimizing the expected cost per unit

of time that includes the sum of the costs of preparation, conservation, and shortage.

The model has 3 hypotheses:

The unmet demand during the delivery time accumulates.

More than one active order is not allowed.

The distribution of demand during the delivery time remains

stationary (without change) over time.

To develop the total cost function per unit of time, let:

f( xProbability
) distribution function of demand x during the time of

delivery.

D Expected demand per unit of time.

12
 Probabilistic Inventory Models

h Storage cost per unit of inventory and per unit of time.

p Cost of stockout per unit of inventory.

K Preparation cost per order.

Based on these definitions, the elements of the function will now be determined.

cost.

1. Preparation cost: The approximate amount of orders per unit of

time is , so the approximate preparation cost per unit of

time is .

2. Expected storage cost: The average inventory is

( y +E {R - x+ })E{R - x} y
I= = + R - E{x}
2 2

The formula is based on the average of the expected initial and final inventories.

of a cycle,y + E R-{ x yE{R

} − x}respectively. As an approximation,

the equation does not take into account the case in whichR - E{x}might be

negative. The expected cost of holding inventory per unit of time is

then equal tohl.

[Link] cost per shortage: There is a shortage when > . Thus, the quantity

∞
expected missing per cycle is = ( ∫− ) ( )∫ How is it supposed to

it is only proportional to the missing quantity, the expected cost of the shortage by

13
 Probabilistic Inventory Models

cycle is , and for cycles per unit of time, the cost of shortage per unit

of time .

The total cost function per unit of time that results is

DK y pD ∞
TCU (and,R
) = + h( + R – Ex) +
{ } ∫ (x - R)∫(x)dx
y 2 y R

The solutions fory ∗ yR∗ optimum are determined with the equations

∂TCU DK h pDS
= − ( +2) − =0
∂y y 2 y2

∂TCU pD ∞
= h −( ∫ )f x dx =( 0)
∂R y R

This is how you get toy∗ = √ 2D(K+pS) (1)

h

∞ hy ∗
∫ R∗ f( x) dx= (2)
pD

In view of thaty ∗ yR∗ cannot be determined in closed forms from the

equations (1) and (2), a numerical algorithm developed by Hadley and Whitin is used.

to determine the solutions. The algorithm converges in a finite amount of

iterations, as long as there is a feasible solution.

ForR = 0 , the last two equations result in, respectively

14
 Probabilistic Inventory Models

2D(K + pE{x})
ŷ = √
h

pD
ỹ=
h

Yesŷ ≥ ỹthere are unique optimal values ofyyRIn the solution method, there

∗ 2DK
recognizes that the minimum value ofyis √, which is reached whenS = 0.
h

The steps of the algorithm are the following:

Step 0: Use the initial solution 1=

∗= √2D, and to do 0= 0Equalizei = 1
ℎ

and continue on the way .

Step 1. Use to determine with equation (2). If ≈ -1, stop;

∗= ∗=
the optimal solution is 1y Otherwise, use in the

equation (1) to calculate Equalizei = i + 1and repeat the step .

Exercise:

Electro uses 1000 gallons of resin per month in the manufacturing process. It costs $

100 place an order for a new batch. The storage cost per gallon and per

my cost is $2 and the cost of the shortfall per gallon is $10. Historical data indicates that

The demand, during the delivery time, is uniform within the interval (0–100).

gallons. Determine the optimal ordering policy for Electro.

Solution:

Data:

15
 Probabilistic Inventory Models

D = 1000 gallons per month

K = $100 per order

h = $2 per gallon and per month

p = $10 per gallon.

f(x) = , ≤ ≤

E(x) = 50 gallons

Check if it has a feasible solution:

The equation of S is calculated:

S is used in equations (1) and (2)

16
 Probabilistic Inventory Models

The last equation results in:

Solution from equations (3) and (4)

Step 1:

Step 2:

17
 Probabilistic Inventory Models

Therefore:

Step 3:

Therefore:

The calculations of R2 and R3 are approximately equal which indicates that the solution

The optimal approximate is: R*=93.61 gallons and y*=319.4 gallons

1.2 SINGLE PERIOD MODEL

Inventory models for a single item are presented when ordering it just once,

to meet the demand in the period. For example, fashion items become

obsolete at the end of the season. Below are two models, which

they represent cases with preparation and without preparation.

18
 Probabilistic Inventory Models

The symbols that will be used are the following:

c = cost of purchase (or production) per unit.

K = Cost of preparation per order.

h = storage cost per unit held in the period.

p = Penalty for missing unit of the period.

D = Random variable representing the demand during the period.

f(D) = Distribution of the probability function of demand during the

period.

y = Quantity ordered.

x = Amount on hand before placing an order.

The model determines the optimal value of y that minimizes the sum of expected costs.

of purchase (or production), storage and shortage. Given the optimal y (= y*) the

Inventory policy establishes to order and* - x if x < y; otherwise no order is placed.

1.2.1 MODEL WITHOUT PREPARATION

The hypotheses of this model are:

demand is presented instantly at the beginning of the period

immediately after the order is received.

No preparation cost is incurred.

19
 Probabilistic Inventory Models

D<y D>y

D y
D
y 0
y-D D-y
Tiempo
0

The figure demonstrates that the inventory position is satisfied after the

demand. If D is less than y, the amount y - D is stored during the period. In case

On the contrary, a deficiency of D–y will be presented if D is greater than y.

The expected cost E{C (y)} for the period is:

y ∞
E{C (y)} = c(y–x) + h y∫ 0- Df(D)dD
( ) +p ∫ y ( D - yf) D
( )dD

It can be proven that the function E{C (y)} has a unique minimum because it is convex.

The first derivative of E{C (y)} with respect to y is taken and set to zero as

continue:

y ∞
c + h f∫ 0D( dD-
) p ∫ y f( D) dD = 0

or well:

c + hP{D ≤ y} –p(1 -P{D ≤ y}) = 0

so:

P{D ≤ y*} =p−c

p+h

20
 Probabilistic Inventory Models

The value of y* is only defined if the critical relation,p−c , it is not negative, that is, if p ≥
p+h

c. The case where p < c makes no sense, because it implies that the purchase cost of

article is greater than the penalty for not providing it.

The previous development assumes that demand D is continuous. If D is discrete, then

f(D) is only defined at discrete points and the cost function is defined as follows:

E{C (y)} = c( y–x) + h ∑yD=o ( y - Df) D+

( ) p ∑∞ D=y+1 ( D - yf(D)
)

The necessary conditions for optimization:

E{C (y -1)} ≥ E{C (y)} y E{C (y + 1)} ≥ E{C (y)}

These conditions are also sufficient in this case because E{C (y)} is a function.

convex. After some algebraic manipulations, the application of those

conditions lead to the following inequalities to determine y*:

P{D≤ y*-1} ≤p−c≤ P{D ≤ y*}

p+h

Exercise:

The owner of a newsstand wants to determine the amount of

newspapers that must be delivered to him daily early in the morning. It takes him 30

cents per copy, and sells it for 75 cents. The sale of newspapers is usually between

7:00 and 8:00 A.M. The newspapers that were not sold by the end of the day are recycled.

still costs 5 cents per copy.

How many copies should they deliver to him each morning assuming that the daily demand

it can be approximated with

21
 Probabilistic Inventory Models

a) A normal distribution with an average of 300 items and standard deviation

of 20 copies.

b) A function of discrete probability distribution as follows:

D 200 220 300 320 340

f(D) 0.1 0.2 0.4 0.2 0.1

The storage and penalty costs are not directly defined in this.

case. However, the problem data indicates that each unsold copy

slope

30–5 = 25 to the owner, and the penalty for terminating it for the newspapers is: 75

-30 = 45 per item. Thus, based on the parameters of the inventory problem, we

you can assume that:

c= 30 per copy.

h = 25 per item.

p = 45 per copy.

First, the critical ratio is determined as follows:

22
 Probabilistic Inventory Models

a) A normal distribution with an average of 300 copies and standard deviation

of 20 copies.

The demand is N(300;20). The standard normal variable is defined as follows:

∗}
−
{ ≤ =
+

According to the standard normal distribution tables

So:

As a consequence, the economic order quantity is y* = 284.2

b) The demand D follows a probability distribution function

discreteF(D). First, the cumulative distribution function P {D is determined

<= y}

23
 Probabilistic Inventory Models

−
P{D≤y* - 1}≤ ≤P{D≤y*}
+

For the critical ratio calculated at 0.214,

P(D≤ 200) ≤ 0.214 ≤ P(D ≤ 220)

As a result, n* = 220 copies

1.2.2. MODEL WITH PREPARATION

Optimal policy in a single-period model with setup cost.

This model differs from the other model in that it incurs a preparation cost K. If

the same notation is used, the expected total cost per period is:

E{Ĉ (y)} = K + E{C(y)}

24
 Probabilistic Inventory Models
y ∞
= K + c(y–x) + h y -∫ 0Df(D)dD
( ) +p ∫ y ( D - yf) D
( )dD

As mentioned earlier, the optimal value y* must satisfy:

P{y ≤ y*} =p−c

p+h

As K is constant, the minimum value of E{Ĉ (y)} must also be presented in y, as

it is seen in the presented figure. The values of s and S shown in the figure will be defined in

brief.

S = y* and the value of s(<S) is determined by the equation

E{Ĉ (s)} = E{Ĉ (S)} = K + E{C(S)}, s < S

(This equation produces another value s1 > S, which is discarded)

If the quantity on hand before placing an order is x units, how much should be paid?

ask? This question is investigated under three conditions:

x<s

s≤x≤S

x>S

Case 1 (x < s). Since x is already at hand, its equivalent cost is E{C(x)}. If it is requested

any additional amount and– x (y > x), the corresponding cost if y is E{Ĉ (y)}, which

it includes the preparation cost K. The figure shows that

min > x E{C(y)} = E{C(S)} < E{C(x)}

Thus, the optimal inventory policy in this case is to order S–x units.

25
 Probabilistic Inventory Models

Case 2 (s ≤ x ≤ S):

E{C(x)} ≤ min>y E{Ĉ(y)} = E{Ĉ(S)}

Therefore, it is not advisable to ask in this case. Consequently, and* = x

Case 3 (x > S): for y > x;

E{C(x)} < E{Ĉ(y)}

This condition indicates that it is not advisable to ask in this case; that is to say y* = x.

The optimal inventory policy, often referred to as the s–S policy, is

summarize as follows:

If x < s, request S–x

If x ≥ s, do not request

The optimality of the s–S policy is guaranteed because the corresponding function of

cost is convex.

Exercise:

The daily demand for an item during a single period is presented in the form

snapshot at the beginning of that period. The probability distribution function of the

demand is uniform, between 0 and 10 units. The unit storage cost of the

The article during the period is $0.50, and the unit cost of the penalty for deficiency

of the same is $4.50. The unit purchase cost is $0.50. A cost is incurred

26
 Probabilistic Inventory Models

fixed at $25 every time an order is placed. Determine the optimal policy of

inventory for that item.

To determine *, it is calculated

− . − .
= = .
+ . + .

Also:

∗
1 y*
{ ≤ } =∫
∗ =
10 10
0

Therefore, S = y* = 8

The expected cost function is

The value of S is calculated by solving

This results in:

27
 Probabilistic Inventory Models

As S = 8, the previous equation reduces to:

The solution to this equation is s = -2 or s = 18. The value of s = 18 (> S) is discarded.

Since the remaining value is negative (= -2), s does not have a feasible value. Therefore, the solution

optimal establishes not to ask (figure 16.7). This conclusion is usually reached when the

cost function is "flat" or when the preparation cost is high in relation to the

other model costs.

1.3 MULTI-PERIOD MODEL

This section describes a model for multiple periods with the hypothesis that not

there is a preparation cost. In addition, the model allows for the accumulation of demand and assumes

zero delays in delivery. It also implies that the demand D in any period is

describe with a stationary probability distribution function, f (D).

The multi-period model takes into account the discounted cash flow of money. If

α (<1) is the discount factor per period, the amount $A available at n

periods counting from today, has a present value of$ αA

28
 Probabilistic Inventory Models

It is assumed that the inventory situation covers n periods, and that in one demand it does not

Satisfied can be accumulated for exactly one period. The following is defined:

Fi (xi) = maximum expected utility for periods i, i + 1, ..., and n, is xi is the

amount on hand before placing an order in period i.

Using the one-period model notation and assuming that r is the income from

unit, the inventory situation can be formulated with the following model of

dynamic programming. i + 1

In (the )house= ( i-2 xi

max{− cy

yi2
+ ∫ rD −[ hyI− Df( D2 dD )] ( )
0

∞
+ ∫ ryi+[ arD ( D- yi 2f) D dD
2 - yI- p ( )] ( )
yi2

∞
+ α ∫ Fi+1 yi2-( D2 f D d(D)}
) ( ) , i = 1,2, … n
0

In whichFn+1 y2n-( D=20. The

) value of xi can be negative, because it can

accumulate the unmet demand. The amount αr (D –yi) is included in the second

integral, because (D–yi) is the unmet demand in period i that is due to

satisfy in the period i + 1.

The problem can be solved recursively. In the case where the quantity of

periods is infinite, the recursive equation reduces to:

29
 Probabilistic Inventory Models

F( xmax{−
) cy - x ( )

y
+ ∫ rD[- hy - D( f D dD
)] ( )
0

∞
[
+ ∫ ry +Dr D - y-( p D )- yf D( d D )] ( )
y

∞
+ α ∫ F y}( ) () ( )
0

Where x and y are the inventory levels for each period, before and after.

receive an order, respectively:

The optimal value of y can be determined from the following necessary condition,

which also turns out to be sufficient, because the expected income function F(x) is

concave.

y ∞
∂(. )
= −c - h∫ f D d( D+
) (∫ 1) − α r+ p[(f D d )D ] () ( )
∂y 0 y

∞ ∂F(y− D)
+ α∫ f( D) d( D=
) 0
0 ∂y

In value of∂F(y−D) It is determined as follows. If there are β (> 0) additional units on hand at
∂y

in the next period, the profit for the next period will increase by cβ, because

we need to ask for that amount less. That means that:

30
 Probabilistic Inventory Models

∂ F(y − D)
=c
∂y

Thus, the necessary condition is:

y y ∞
− c- h f ∫D0 d( D+
) ( 1) - α)r[ + p(1 − f D] d(D)+αcf
∫ 0 ( D) d D) = 0 ∫ 0 ( ) ( )

Then, the optimal inventory level is determined with:

y
p +1(+ α (r )-c)
∫ f D( d) D( =)
0 p + h +1 (- α r )

The optimal inventory policy for each period, given its initial inventory level x

it is then:

If x < y*, ask for y* - x

If x ≥ y*, do not ask

EXERCISES

1. Regarding determining the neon light inventory policy, it has been determined that the

Economic order quantity of CEP = 100 units. If the daily demand is

normal, with an average D = 100 lights and a standard deviation σ = 10 lights, that is, N

(100, 10), determine the size of the reserve such that the probability of it running out

the existence is less than ∞ = 0.05. the effective delay time is L = 2 days.

SOLUTION

microliter 200

31
 Probabilistic Inventory Models

= DL = 100∗ 2

=√ 2
= √ 1022

σL = 14.14 Units
According to the normal distribution tables K0.05 = 1.645. Consequently, we

calculate the size of the reserve as follows:

B ≥ 14.14∗ 1.645 23 Neon Lights

As the CEP = 1000 lights, politics optimum of

inventory with a reserve B establishes purchasing 1000 units whenever the level

I lowered the inventory to 223 (=K + µL = 23 + 2 * 100) units.

+ = 23 +2* 100 = 223

A newspaper vendor pays $0.20 for each newspaper and sells it for $0.50.

therefore:

$0.20 0.3 P
= 0.60
$0.30 0.5 =

What is the probability that the newspapers sell, now it is interesting to determine the

quantity Q of newspapers to buy, for which the value of Z is determined in the

normal distribution.

SOLUTION

32
 Probabilistic Inventory Models

= . NORM. ESTAN. INV. (0.6)

Z
0.2533471
=

Therefore, if based on historical data the average number of newspapers sold is 90 with

standard deviation of 10, the number of newspapers to buy is the mean plus one

extra amount:

Sales 90

Desv. Stand 10

X = 90 + 0.2533471 *10

X= 93
The extra purchase quantity is 3

= 90

3. The selling price of a product is $100 per unit and has a constant cost of

$70 per unit. Each unsold unit has a salvage value of $20. It

wait for the demand to be between 35 and 40 units per period. Its

the probabilities are as follows:

Demand Probability of the Probability

Demand Accumulated
35 0.10 0.10
36 0.15 0.25
37 0.25 0.50
38 0.25 0.75
39 0.10 0.90
40 0.10 1.00

33
 Probabilistic Inventory Models

SOLUTION

How many units should be ordered?

100−70 Cu
$30
= Underestimate

= 70− 20 CO
$50
= Overestimate

30
≤
30 + 50

P≤ 0.38

This corresponds to 38 units

4. The company Sharp Inc. is a company that sells hypodermic needles.

painless in hospitals, wants to reduce its inventory costs by means of the

determination of the number of needles that must be obtained in each order. The demand

annual is 1000 units; the handling cost per unit per year is 0.50

dollars. Calculate the optimal number of units per order.

SOLUTION

DATA:

D= 1000 Annual demand

S/.
A
10 Cost to charge an order
S/
H=
0.5 /Unit - year (Cost per unit to maintain in inventory per year)

a) Optimal number of units per order.

2 2(10)(1000)
∗= √ ∗= √
0.05 34
 Probabilistic Inventory Models

Q* = 632 Needles/order

b) Number of orders in the year (N)

1000
= =
632

N= 2 orders

c) Cycle time, expected time between orders (T)

632
= =
1000

T= 0.6

= 20 days/order

The year is considered = 652 working days

= 2 Units/days

d) Reorder point based on inventory

Delivery time = 20 days

When there are 20 units in inventory, the following order has to be made.

e) Annual Variable Costs of the Inventory System (VC)

1000 632
= 10 ( ) + 0.5( )
632 2

Vc = 174

35
 Probabilistic Inventory Models

5. PettersonElectronic supplies electronic circuits to a company that installs

microprocessors in refrigerators and other white goods. One of the

the component has an annual demand of 250 units, which remains constant

throughout the year. Inventory maintenance is estimated at $1 annually for

The unit cost for placing the order is $20 per order.

a) To minimize the cost, how many units should be ordered each time

Is an order placed?

b) How many annual orders are needed with the optimal policy?

c) What is the average inventory if costs are minimized?

d) Suppose that the cost of placing the order is not $20, and Patterson has

I have made several orders of 150 units. For this order policy to be

optimal. What should the cost be to place the order?

SOLUTION

a. The assumptions of EQQ are met in such a way that the optimal order quantity is

∗=
2 0
= √

2 (25020
)
= ∗= √
1

EOQ =
100
Q* = Units

b. Number of orders per year

36
 Probabilistic Inventory Models

X orders by
2.5
= year

Note that this amount would mean that in a year the company places 3 orders and

in the following I would only need to place 2, since the inventory would be maintained from

of the order from the previous year. This averages 2.5 orders per year.

c. Average Inventory

= 50 units
2

d. Considering that there is an annual demand of 250 units, and a cost per

inventory maintenance of $1, and an order quantity of 150, Patterson

Electronics must determine what the cost of placing an order would be for the

the order policy of 150 units is optimal. To find the answer to this

problem, it is necessary to solve the traditional EOQ equation to calculate the cost of

place the order. As can be seen in the calculations presented at

continuation, a cost of $45 is required for the order quantity that

correspond to 150 units being optimal.

2
2 0 0=
= √ 2

C0 = 45 Dollars
(150)2(1)
0=
500

37
 Probabilistic Inventory Models

Fleming Accessories produces paper shredders used in offices and in

art stores. The mini shredder is one of their most popular items: the demand

constant annual throughout the year is 6750 units. Kristen Flemming,

the owner of the firm produces mini paper shredders in batches. On average,

I can manufacture 125 mini units per day. The demand for these devices during the

The production process amounts to 30 per day. The cost of the startup of

The necessary equipment for the production of the mini crushers is $150. The costs of

maintenance amounts to $1 per year per device. How many mini shredders should I have?

produce Kristen in each batch?

SOLUTION:

The data from FlemmingAccessories is summarized as follows:

D= 6750 Units
Cs = $150.00
Ch = $ 1.00
d= 30 Units
p= 125 Units

This is a production problem that involves the daily production rate.

as the daily demand rate. The appropriate calculations are shown below:

∗=
2 2(6750)(150)
∗=
√ √ 30
(1 - ) 1(1 −125 )

Q* = 1632

38
 Probabilistic Inventory Models

7. DorseyDitributors has an annual demand of 1400 metal detectors. The cost

The typical cost of a Dorey detector is $400. The inventory maintenance costs

it is estimated at 20% of the cost of the unit, while the unit cost for performing a

the order is $25. If Dorsey places an order for quantities of 300 or more, he could

obtain a 5% discount on the cost of the detectors. Should Dorsey accept the

volume discount? Consider that the demand is constant.

SOLUTION:

The solution to any volume discount model involves the determination

of the total cost of each alternative after the quantities have been calculated and

Adjusted for the original problem and for each discount. The analysis begins without

any discount.

D = 1400
C0 = $25
I= 0.2
C= $400

2 0
=√

( )
2(1400)(25)
= √
0.2(400)
EOQ = 29.6 Units

Total Cost (Without discount) = Material Cost + Cost for

place the order + maintenance cost

(
1400($25)29.6 $400)(0. 2) )
( ) = $400(1400) + +
29.6 2

39
 Probabilistic Inventory Models

CT (Sin descto) $562,366

The next step is to calculate the total discount cost:

C = $380

Due to the fact that this last order amount is below the discounted price,

we need to adjust the order quantity to 300 units. The next step is to calculate

the total cost:

( )
2(1400)25)
= √
0.2($380)

Total Cost (With discount) = Cost of Material + Cost for carrying out the
Order + Maintenance cost

(
1400($2300$380)(0.2 )
( ) = $380(1400) + +
300 2

CT (Without Deduction) = $543,517

The optimal strategy is to place an order of 300 units for a total cost of

$543,517

40