Inventory Management Models

Single-Period Model (the newsboy model)

Co = cost/unit for overestimating demand
Cu = cost/unit for underestimating demand
P = the probability that the unit will be sold
P < Cu/(Co+Cu)
Example: Your group wants to sell bags of popcorn at the homecoming football game !ata from the last "#
homecoming games reveals that sales have averaged $## bags of popcorn with a standard deviation of %# bags
&he popcorn is bought from a local vendor who does not accept returns of unsold popcorn Your cost is '#((/
bag and you sell the popcorn for '"##/bag )ow many bags should your group buy to sell at the game*
Co = '#(( Cu = '#+( P = '#+(/('#((+'#+() = #+(
&he #+( translates to the service level, &herefore- loo. up #+( in the body of a cumulative standard normal
table to find the appropriate / vale 0f course- since P<#(#- the / value will be a negative number,
/ = 1"%2 &herefore- you would order $## + (1"%2) 3 %# = %4+5 =6 %4+
Multi-Period Models
Basic EOQ Assumptions
" !emand is uniform and .nown (d- !)
% &he item cost is fi3ed (P)
$ &he order cost is fi3ed (7)
+ &he holding cost is fi3ed ()- .P)
( &he lead time is constant (8)
2 0rders are received in completed form (no split shipments)
"
!wo Basic "n#entory Management $ecisions
" )ow much should be ordered each time an order is placed (9)
% :hen should an order be placed (;0P)
Simple EOQ Model
9< = (%!7)/) or 9< = (%!7)/(.P)
&C< = (!/9<)7 + (9</%)) or &C< = (!/9<)7 + (9</%).P
;0P = d8 where d = !/& (& is the number of days or wee.s open for business
depending on the time unit for lead time)
Simple EOQ Model with Sa%ety Stoc&
=f we rela3 assumptions " and/or (- then we need to add a safety stoc. (77) component to the ;0P>
;0P = d8 + /8d
%
+ d
%
8
%
where /8d
%
+ d
%
8
%
= 77 and / represents the service level
&C< = (!/9<)7 + (9</%)) + 77 3 ) or &C< = (!/9<)7 + (9</%).P + 77 3 .P
Example: ? product has a daily demand of (# units on average with a standard deviation of "% units &he lead
time averages $ days with a standard deviation of " day 0rdering cost is '"##- holding cost is '%/unit/year- and
the business is open %(# days/year =f we desire a @(A service level- then how much should we order each time
we place an order and when should those orders be placed*
9
<
= [%((#3%(#)"##B/% = "-""5
;0P = ((#)($) + "2(($)("%)
%
+ ((#)
%
(")
%
= %+#
Production 'un Model
=f we go bac. to the original set of assumptions and then rela3 only assumption 2- then we need to
compensate for split deliveries>
9< = (%!7)/())("1(!/;)) or 9< = (%!7)/(.P)("1(!/;)) or 9< = (%!7)/())("1(d/r)) or
9< = (%!7)/(.P)("1(d/r)) where ; is the annual production rate- r = ;/&- and ;6!
&C = (!/9<)7 + (9</%)())("1(d/r)) or &C = (!/9<)7 + (9</%)(.P)("1(d/r))
(Cote> &he ratios !/; and d/r are identical,)
%
Example: ? firm assembles (## units of a product each day in the final assembly department =n one of the
fabrication departments- a component used in the product can be produced at a rate of 4(#/day ?nnual demand
for the product is "%(-### units- the setup cost is '+##- and the holding cost is '"#/unit/year )ow many units
of the component should the final assembly department order each time it places an order*
9
<
= (%3"%(-###3'+##)/D'"#("1(##/4(#)B = (-+44%$ =6 (-+44
&C = ("%(-###/(-+44)'+## + D((-+44("1(##/4(#))/%B'"# = '@-"%@#@ + '@-"%5$$
Quantity $iscount Models
=f we go bac. to the original set of assumptions and then rela3 only assumption %- then we need to
compensate for price differences &o ma.e the compensation- we use an algorithm that allows us to
ma.e price comparisons using the steps listed below>
" Eind the feasible F09 and calculate itGs total cost where
&C = (!/9<)7 + (9</%)) + P! or
&C = (!/9<)7 + (9</%).P + P! where . = percent holding cost
% =f the feasible F09 is associated with the lowest cost curve- then you are finished,
=f not- go to step $
$ Calculate total costs for each price brea.
+ Choose the order Huantity associated with the lowest total cost
Case I: Constant Holding Cost
9< = (%!7)/) Eor Case =- there is only one F09 calculation since 9< is not a function of price
&his satisfies step " of the algorithm
Case II: Variable Holding Cost
9< = (%!7)/(.P) Eor Case ==- begin calculating 9< using the lowest value for P and continue to
wor. up the price schedule until a feasible value for 9< is found &his satisfies
step " of the algorithm
Example (Case I versus Case II): &he following price schedule is in effect )olding costs are (=) '(/unit/year
or (==) $#A of the purchase price- order cost is '%##- and annual demand is estimated to be "#-### units
Eor each case- what is the least e3pense order plan*
$
9uantity Price/Init
"1+@@ '"###
(##14@@ ' @5#
5##1@@@ ' @4(
"### or more ' @2(
Periodic Order Quantity Model
:e now move away from the simple F09 model and consider the situation where we are constrained in
delivery schedule by a vendor or some other force =nstead of using a reorder point- we periodically
count inventory and then determine how much we have to order to replenish our stoc. &he periods
are fi3ed
9< = d(8+;P) + /JdK8+;P L =
:here ;P (;eorder Period) = the number of days- wee.s- etc between deliveries and
= = the current inventory level
Example: ? grocery store that is open 2 days a wee. receives deliveries every $ days from one of itGs suppliers
? particular product has a daily demand of "(# units with a standard deviation of %# units &he lead time is
fi3ed at % days Currently there are +## units of this product on the shelf and itGs time to call in an order &he
grocery store li.es to maintain a @(A service level for this item )ow much should be ordered*
P = $ 8 = % = = +## / = "2(
9 = "(#(%+$) + "2((%#)%+$ 1 +## = 4(# + 4+ L +## = +%+
+
( !ables
(
2