Some Comments on 'Inventory Models with Fixed and Variable Lead Time Crash Costs

Consideration'
Author(s): M. A. Hoque and S. K. Goyal
Source: The Journal of the Operational Research Society, Vol. 55, No. 6 (Jun., 2004), pp. 674-
676
Published by: Palgrave Macmillan Journals on behalf of the Operational Research Society
Stable URL: http://www.jstor.org/stable/4101974
Accessed: 21/11/2010 01:18

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=pal.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Palgrave Macmillan Journals and Operational Research Society are collaborating with JSTOR to digitize,
preserve and extend access to The Journal of the Operational Research Society.

http://www.jstor.org
Journal of the Operational Research Society (2004) 55, 674-676 ? 2004 OperationalResearchSociety Ltd.All rightsreserved. 0160-5682/04 $30.00
www.palgrave-journals.com/jors

Viewpoint

Some commentson 'Inventorymodelswith fixed and variable formulatedthe expectedannual cost, EAC(Q, L), as
lead time crashcosts consideration'

Journal of the OperationalResearch Society (2004) 55, 674-676.

doi:10.1057/palgrave.jors.2601762
EAC(Q,L)
Q
=A-D
2
+h[

-
- Q
+ kav + (1-

+ [-E + ro(1 /)]a]V-L'P(k)

)avoYL(k)

Q
+ D ai(Li-- L)+ a,(Tj- tj)
Introduction
We would like to comment on the paper by Pan et al,1 in - L) + -
+ D bi(Li-
which assuming reduced lead time crashing cost as a bj(Tj1 tj)
function of both the order quantity and the reduced lead t- -
(1)
time, they presented two inventory models, the one with
known deterministicdemand and the other with unknown where TP(k) = 0b(k)-k[1-(D(k)] and Li= E= 1T-
demanddistribution.The solution proceduresare illustrated Z-I'(Ti-tj) for Li<L <Li-1. For correct calculation of
with numericalexamples. We have dealt with the determi- the total cost bindingson L should be Li< L < Li1.
nistic demand case only. This viewpoint highlights For a fixed Q, they have shown that the minimum
the misleading behaviour of the formulae used to obtain expectedtotal annual cost occurs at one of the endpointsof
the optimal order quantity. It also demonstrates the the lead time intervals.Setting OEAC(Q,L)/6Q to zero and
infeasibilityof the model due to the lack of a constrainton solving for Q, they obtained
the orderquantityin orderto satisfy the demandin the lead
time. The model is extendedwith the additionof a constraint + [7r+ to(1
to satisfy the demand in the lead time. An optimal solution Q[•2D-{A -fl)]a/LT(k)
technique of the extended model is presented, and a i--1/2 (2)
comparativestudy of the results of the numericalexample +ai(Li-I-L) + -
is carriedout. j= I)
aj(Tj tJ)

The minimum and maximum of the optimal order

Analysis of the model quantityin Equations(3) and (4) of theirpaper,respectively,
are shown as
They presentedtheir models using the following notations: - 1/2
L the length of lead time (in weeks); Q= order quantity;
D = average demand per year; A = fixed orderingcost per Qmin 2D A + [7n+
?o(l - #)]aP(k) 1t
order;h = inventory holding cost per unit per year; r = the J=1

reorder point; rno=the gross marginal profit per unit;

7[ = fixedpenaltycost per unit short; 1 = the averagedemand when none of the lead time components is crashedat all,
rate in units per day; a= the standard deviation of the
demand rate; k=safety factor; =the standard normal Qmax A + [ + -
distribution;q= the standard normal cumulativedistribu- -•2D
tion function; P= the backorder ratio; n= the number of ~-o(1 fl)]-•P(k)
mutuallyindependentcomponentsof the lead time; Ti= the T + Ea(T, -
normal duration of the ith component; ti the minimum tt)
durationof the ith component;Li = the lead time length (in x?
weeks) with component i (i= 1,...,n), crashed to their when all the lead time components are crashed to their
minimum values; ai=unit fixed crash cost per week; minimumlimits.
and bi= unit variable crash cost per week for the ith These formulaeare not correctbecauseif none of the lead
componentof the leadtimereduced.Withthesenotationsthey time componentsis crashed,the lead time is and
" T1
 675
Viewpoint

in case of crashing all the lead time components to their
minimumlimits, the lead time is Therefore,in the
,/ , tj
above formulae these mathematical expressions must be
interchanged.Thus they developedtheir algorithmbased on
erroneous formulae providing misleading solution techni-
ques. Besides, they developed their solution technique
without imposing any constraint on the order quantity in
orderto satisfythe demandin the lead time. For this reason,
their solution technique has given infeasible solutions for
different values of / for the numerical example when
demand D= 5500. These optimal solutions are shown in
Table 1.
Note that the demandper week is 110 and hence none of
the optimallot sizesmeets the demandfor the corresponding
optimal lead time. Thus the solution techniqueleads to an
infeasiblesolution.
An alternative solution technique
In order to satisfy the demand in the lead time, the order
quantity must be greater than or equal to the required
demandin the lead time, that is,
where Q?is the value of Q calculatedfrom Equation(2). The
total cost calculatedfrom (1) for Q= Q* with Li< L < Li- is
the corresponding minimal cost. The minimum of the
minimal costs thus calculated for all i gives the final
minimum cost. The value of Q associated with the final
minimum cost is the minimal order quantity. Following
this solution approach, the numericalexample 1 solved by
Pan et all is solved and comparativeresults are given in
Table 2 (in the table, PHL denotesthe method developedby
Pan, Hsiao and Lee and HG denotes the method developed
by us):
For the same lot size and lead time, the total minimal
costs calculated following our approach are always found
to be different from the correspondingone they put in
their Table 4. These may be due to the limitation
Li<L<~Li_1 in their formulation of the total cost.
For D=5500, the lot sizes found out for minimal
total costs by Pan et al's method do not satisfy the
demand for the lead time and hence the solutions are
infeasible, whereas our approach provides minimal cost
solutions satisfyingthe demandin the lead time plus safety
stock.

Q~ DL/50 + kav-L (assuming50 weeksin a year) (3)

Concluding remarks
This viewpoint highlights the erroneous formulae pro-
For each value of i, the optimalvalue of Q denotedby Q*
posed by Pan et al. It also demonstratesthat the model
can be determinedas follows:
proposed by them leads to an infeasible solution because
of the lack of constraint on the order quantity in order
Q* = Max(Q0, DL/50 + ka/L) (4) to satisfy the demand in the lead time plus safety shock.
To satisfy the demand in the lead time, the model is
extended to include a constraint on the order quantity.
An optimal solution technique of the extended model
Table1 Infeasibleoptimallot sizesof the example1 for is proposed, which is able to provide minimal cost
differentvaluesof P whenD = 5500 solution of the problem considering feasibility. The
Valueof # 0.0 0.5 0.8 1.0 potentiality of our method is shown by a comparative
Optimallot size 573 510 454 412 study of the results of the numerical example they
Optimallead time (in weeks) 6 8 8 8 illustrated.

Table2 Comparativeresultsof exampleI with deterministic

demandfor differentvaluesof P
Values of / Solution by D = 250 units/year D = 600 units/year D = 5500 units/year

Q, L EAC(Q, L) Q, L EAC(Q, L) Q, L EAC(Q, L)

0.0 PHL 115,3 2587.81 178,3 3931.24 573,6 12087.29a
HG 115,3 2591.50 178,3 3936.88 542,4 12335.06
0.5 PHL 103,3 2344.99 158,4 3553.96 510,8 10554.44a
HG 104,3 2247.52 159,4 3558.49 480,4 11087.741
0.8 PHL 93,4 2165.93 145,4 3269.75 454,8 9420.76a
HG 93,4 2167.99 145,4 3272.89 452,4 10246.04
1.0 PHL 87,4 2031.39 136,8 3058.24 412,8 8580.80a
HG 87,4 2032.74 136,8 3061.17 452,4 9673.22
alInfeasiblesolutionbecausethe lot size does not satisfythe demandfor the lead timeplus safetystock.
676 Journal
oftheOperational
Research Vol.55, No.6
Society

Acknowledgements--We acknowledge that the authors of the original University Sains Malaysia MA Hoque
paper pointed out a typo in Equation (1) and the authors of the view
point are grateful to them.
and
Concordia University, SK Goyal
Canada

References
Editor's note: The authors of the orginal paper have seen
1 Pan JC-H,Hsiao Y-C and Lee C-J (2002).Inventorymodelswith
fixed and variablelead time crashcosts considerations.J OplRes
and agreed to the publication of this viewpoint.
Soc 53: 1048-1053. T Williams