ARTICLE IN PRESS

Int. J. Production Economics 93–94 (2005) 345–356

www.elsevier.com/locate/dsw

Stability of perishable goods in cold logistic chains

Marija Bogataj, Ludvik Bogataj, Robert Vodopivec
Faculty of Economics, University of Ljubljana, Kardeljeva ploščad 17, 1000 Ljubljana, Slovenia

Abstract

We consider that production and logistics facilities in the supply chain are positioned between the origin and the
supply market or in a part of it. Any changes in time-distance or temperature in the chain could cause the net present
value of the activities and their added value in the supply chain to be perturbed. In reality the perturbations can be
robust. The natural question arises is to what are the effects of some perturbations in a supply chain, in its production
or distribution part, on the stability of perishable goods in such systems and what is the appropriate control which
keeps the product on the required level of quality and quantity at the final delivery. These analyses are especially
important to assure the stability of cold chains in the cold chains management (CCM). What conditions should be
fulfilled to assure that after such robust perturbations of parameters and especially time delays (lead times and some
other delays) the behaviour of the logistics chain would still be within the prescribed limits. The formulation obtained in
the time domain will be compared with the formulation in the frequency space introduced by Grubbström in his paper
‘‘On the Application of the Laplace Transform to Certain Economic Problems’’ (Manage. Sci. (7) (1967) 558) and with
the later studies of his school. Their model was extended by the authors to the complete logistic chain, where location —
and with it also the distance between the activity cells of logistic chain — play an important role.
r 2004 Elsevier B.V. All rights reserved.

Keywords: Stability; Logistic chain; Lead time; Input–output analysis; Laplace transforms; MRP; Robust perturbations; Cold chains;
Cold chains management; Perishable goods

1. Introduction to the global logistics management of cold chains

At the end of the past millennium supply networks management emerges as an important management
paradigm and is said to be one of the last unexplored frontiers of the management science (Baldiwala,
2001). In the global marketplace, getting the right goods at the right time at the right place is as important
as low costs. It is essential for enterprises to develop the visibility of flows in the network and to understand
management tools needed to supply network logistics. Visibility is especially important in cold chains and
other chains of perishable goods where control of changing the temperature in transportation or even in
Corresponding author. Tel.: +386-7-3461570; fax: +386-7-3461571
E-mail address: [email protected] (M. Bogataj).

0925-5273/$ - see front matter r 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpe.2004.06.032
 ARTICLE IN PRESS

346 M. Bogataj et al. / Int. J. Production Economics 93–94 (2005) 345–356

production cells or control by adding the preservatives is needed to keep the quality and quantity of
product at the end of the supply chain on the required level.
What is the global cold chains management (CCM)? It is the process of planning, implementing and
controlling efficient, effective flow and storage of perishable goods, related services and information from
one or more points of origin to the points of production, distribution and consumptions in order to meet
customers’ requirements on a worldwide scale. It is the process of integrating the existing business activities,
including special activities for perishable goods conservation along the value chains, where more suppliers
of certain raw materials or more production cells of certain semi-products appear in order to create value
for the end user.
The global market of refrigerated products and prepared meals is increasing rapidly due to the main
drivers of globalization:
decreasing tariffs (WTO, GATT),
permanent improvement of transportation efficiency (the emergence of air travel, induced traffic on the
new highways, faster ocean crossings, etc.),
development of communication and information technology,
and development of cold chains techniques.
As the market for refrigerated products and prepared meals is increasing, regulations of supply chain
must be prescribed by public authorities to protect final customers from health hazards. The temperature
and other types of preservation control at each stage of the supply chain are essential to maintain the
prescribed quality of the product until it reaches final consumer. This quality is influenced by the perturbed
time delays in actions and by temperature disturbances. Europe and other developed countries have
established a set of regulations for temperature control and equipment performance at different steps of the
cold chain. Risk assessment is necessary and it requires permanent control of the products in the cold chain.
Temperature variations may result in the growth of different pathogen particles as micro-bacteria, and
other deteriorations. Their quantity has to be limited. Since temperature abuse in the cold chain can cause
microbial hazard and losses of product quality, temperature control is essential to keep the final consumer
safe. Many countries have established food safety regulations as:
product temperature regulation along the supply chain,
obligatory recording of air and product temperature in refrigerated vehicles, production cells and
loading-reloading places,
and standardized equipment confirmed by attesting.
As the regulations differ from country to country, European regulations and directives have been
harmonized for prescribed temperature of frozen products at most steps of the production and distribution
chain. The European Community directive EN 92/1 CEE has prescribed air temperature recording in
transportation, warehousing and retail display. We have ‘‘Agreement on the international carriage of
perishable foodstuffs and special equipment to be used for such carriage’’ (ATP) established in 1970, by
which contracting parties have to take the measures necessary to ensure that equipment, used for carriage
of perishable goods meets the technical standards described in the agreement. The standards prescribed by
ATP include limits in heat loss of products ðin W=m2 KÞ and the ratio between the refrigerating unit
capacity and the heat loss of the product. For display cabinets, standards have been adopted
internationally, but are not compulsory everywhere.
Up to consumption, the distribution chain of food products is made up of many different stages. Product
temperature, especially when containers of this chain are not the same from the very beginning, fluctuates much
more than when transportation is provided in the same container with cooling control on all the time. Product
temperature varies in each step, especially during loading and unloading outside air-conditioned warehouses
and can exceed the upper or lower limits, although such circumstances may last only a very short time.
 Recently, the concept of ‘‘cold traceability’’ was introduced, which requires some tools and equipment
such as thermometers, temperature recorders, temperature indicators and time-temperature integrators
mostly for static approach to the quality control of perishable goods. This concept helps to trace different
groups of perishable goods like poultry and other meat, fish, fruits and vegetable, confectionery, ice cream
and other dairy products which are transported under different cooling requirements. The question arises,
till when and how the cargo handling can be improved in order to meet the final requirements? What should
be the conditions and restrictions of a dynamic system management so that the final consumer’s need for
healthy food will be met? Are the static limitations of the perturbations at each moment as good as dynamic
restrictions? To overview the chain flows and to achieve the optimal control in the procedures of cooling
and improvements, input–output approach studied by Grubbström especially in his papers Grubbström
(1967, 1996a, b, 1998), Grubbström and Bogataj (1998), is extended to the entire cold chain as is suggested
by Bogataj and Bogataj (2001) with additional evaluation of deterioration rate in the distance functions.
Here the level and uncertainty of demand, which is met by delivery service level b, is the subject of quality,
mill pricing and location (distance between two activity cells) as is presented in Bogataj (1994, 1995, 1996,
1998, 1999) and Girlich (1995). The results of cold chain management (CCM) are finally expressed and
valuated as the net present value (NPV) of the final delivery reduced by the costs of production,
distribution, additional cooling and deterioration of goods.
 In the case of perishable goods in our supply chain at any of n stages (k ¼ 1; 2; . . . ; n), the development of
the state of the system is described by a set of first-order linear differential-delay equations (DDE):
DDE (4) has the following compact presentation:
_
_ _ _ _ _ _


X ðtÞ ¼ A X ðtÞ þ ðI H Þ PðtÞ þ BðtÞ X ðt; ’n
tG?OpOLp-wfifiG‘p-pT+uLt2TjuA?’pTpTfu-wCTGCp-pT+uLtvTjuA?GT+uC¨fu¨w
’J
 ARTICLE IN PRESS

350 M. Bogataj et al. / Int. J. Production Economics 93–94 (2005) 345–356

Linear DDE (6) can be written in condensed form as follows:

ð6aÞ

or

_
_ _ _ _ _





X ðtÞ ¼ ðA þ DAÞ X ðtÞ þ ðI H Þ PðtÞ þ ðBðtÞ þ DBðtÞÞ X ðt; t1 ; t2 ; . . . tn Þ F ðtÞ; ð6bÞ

where

ð7Þ
 ARTICLE IN PRESS

M. Bogataj et al. / Int. J. Production Economics 93–94 (2005) 345–356 351

In system (6) mass preservation condition (5a),

_ _
G X ðtÞ ¼ 0; t 2 ½0; tf ;

the initial condition (5b),

_ _
X 0 ðyÞ ¼ FðyÞ; maxftk gpyp0;
k

xk0 ðyÞ ¼ ~
and also the sequential conditions (5c): ~ xk 1f ðyÞ, k ¼ 1; 2; . . . ; n,

tk pyp0;
have to be fulfilled.

3. Stabilization of product deterioration processes by cooling

_


The feedback control ðB þ DBÞ X ðt; t1; t2 ; . . . ; tn Þ is needed as additional cooling so that one of two
optimization criterions (a) minimal costs or (b) maximal annuity stream would be fulfilled.
(a) The total cost of additional cooling and the cost of deteriorated portion of cargo, described in (8)
should be minimal. In criterion function J, these costs appear often in quadratic form, where the first
part presents the costs of cooling all over the time, described by matrix C of cooling costs on all stages
of logistic chain, through matrix W, which could be identity. The second part of criterion function
requires that the shortage of good part of total inventory at the point of demand (usually the final stage
of supply chain) is penalized. Also the total inventory, which exceeds the total demand, causes costs of
additional deterioration and additional storage expenses, which also has to be penalized. This is
described
_ by quadratic form OT V O. Here b presents the probability of shortage of goods
(Prob F ðt; bÞoDðtÞ ¼ bÞ and
Z tf T

_ _ _ _ _ _


J¼ S ðtÞðB þ DBÞT C T WCðB þ DBÞ S ðtÞ dt þ ðS ðtÞ F ðt; bÞÞT OT V OðSðtÞ F ðt; bÞÞ dt ; ð8Þ
0

_ _
_
SðtÞ ¼ X ðtÞ: ð9Þ
_
We are looking for such a control U ðtÞ
_ _


~ðt;~
U ðtÞ ¼ ðB þ DBÞ X tÞ ð10Þ
T
that J would be minimal. Here O V O is the weight matrix sometimes having only one nonzero element
on the diagonal at 3n 1 position. The rate OT V O=C T WC presents the degree of stabilization of
perishable goods.
For the solution of this problem, see Manitius (1974).
or
(b) In the case that we consider the net present value of revenues achieved when the good part of products
is sold, reduced for the costs of cooling and other costs in the supply chain, distributed on the time
horizon, when interest rate of costs and revenues is substantial, the approach in frequency space
suggested in general by Grubbström (1967, 1996a, b,1998), Grubbström and Bogataj (1998) gives more
straightforward results.
Let us write Eq. (6b) in frequency space using the following Laplace transform expressions:
Z 1
_ ~
_ _
dfX ðtÞg ¼ X ðsÞ ¼ e st
X ðtÞ dt; ð11Þ
0

 
d _ ~
_ _
d X ðtÞ ¼ sX ðsÞ X ð0Þ; ð12Þ
dt

Z t_ 
~
_
d SðaÞ da ¼ SðsÞ=s; ð13Þ
0

_
Z t_
~
_
dfX ðt;~
tÞg ¼ t ðsÞ d SðaÞ da
0
is placed only at ð3n 1; 3n 1Þ.

ð18Þ

~
_
and Dt
 ARTICLE IN PRESS

354 M. Bogataj et al. / Int. J. Production Economics 93–94 (2005) 345–356

(containers) per TU. The trucks are immediately sent to the next activity cell—Warehouse 2 (W2). They
need 9 hours to reach W2. The temperature here is changed to Temp2. Deterioration rate here is q2 ¼ 0:005
per TU, if there is no additional cooling (control). At the final place, the containers are reloaded and stored
in the warehouse W2, packed in the palettes of 10 boxes. The palettes are waiting for customers. 750K1
palettes per TU are delivered from the currant arrivals and others from the reserves, if exist. Reloading and
waiting last for 4 hours with a constant deterioration rate. Deterioration rate in this warehouse (W2) at a
prescribed temperature Temp3 is q3 ¼ 0:003 per TU. PP has quality control at the next loading (final
delivery F).
The purchasing price at the first stage (input to the first activity cell) is 1000 EUR per box. Selling price is
33 000 EUR per palette of goods PP. Expected cost of shortage of one palette is 2500 EUR. The quantity
which exceeds the normal demand, D ¼ 1000, is sold for 500 EUR per palette.
The additional cooling can be started only by time delay (1, 3, 2) TU sequentially (see ~ tÞ at each activity
cell. Let us assume that we have already decided additional cooling which retards deterioration by the rate
0.002, 0.004 and 0.002, which costs C B ¼ ð 1; 100; 50Þ per unit of PP per unit of cooling rate,
respectively. Let us study how the perturbations of cooling, being of different volume DB which costs
additionally ( 1, 100, 50) per unit of PP per cooling rate, influence NPV! Let us note that for the sake of
simplicity, here the supply chain is very simple; no production of different inputs any particular stage exists,
there are no feedback flows or other properties which would make the matrix H more or less sparse. If the
daily s ¼ ðinterest þ insuranceÞrate ¼ 0:1 our vectors and matrices are the following:

DA ¼ 0;
 ARTICLE IN PRESS

M. Bogataj et al. / Int. J. Production Economics 93–94 (2005) 345–356 355

Table 1
Costs in a numerical example of the supply chain

Stage 3 2 1 3 Good part Shortage Sale NPV

Purchasing costs 0 0 1000 3E+07

Packaging, loading, 440625
transportation and other
activity costs 5 200 10
Additional cooling costs 1 100 50 Variable, approximately 2E+06
Delivery revenue (costs) 0 0 0 33000 2500 500 Variable
Total costs per unit 6 300 1060 33000 2500 500 Variable

Table 2
3
The revenue at the final stage and net present value of total supply chain B=10 (2.1 4.1 2.1)

DB B þ DB ¼ 0 0.0001 0.00001 0.000001 0.0000001

Net final revenue 943962 1.8E+07 3.3E+07 3.3E+07 3.7E+07
NPV 3E+07 1E+07 990750 990750 4734279
DB 0.0000000 0.00010 0.00100 0.00515 0.005155
Net final revenue 3.7E+07 3.7E+07 3.7E+07 3.3E+07 1.8E+07
NPV 4734279 4734273 4733975 990750 1E+07

2 3 2 3 2 3 2 3
1:5 1125 750 0
6 9 7 6 675 7 6 75 7 607
6 7 6 7 6 7 6 7
6 7 6 7 6 7 6 7
6 4 7 6 30 000 7 6 7500 7 607
6 7 6 7 6 7 6 7
6 7 6 7 6 7 6 7
6 1:5 7 6 1125 7 6 K 1 750 7 617
6 7 _ 6 7 6 7 6 7
t¼6 7
6 1:5 7; X ¼ ð0Þ6
6 0 7;
7 F ðsÞ ¼ 1=s6
6 ð1 K 1 Þ 750 7;
7 t¼6
~ 7
6 1 7:
6 7 6 7 6 7 6 7
6 9 7 6 675 7 6 K 2 75 7 637
6 7 6 7 6 7 6 7
6 9 7 6 0 7 6 ð1 K 2 Þ 75 7 637
6 7 6 7 6 7 6 7
6 7 6 7 6 7 6 7
4 4 5 4 30 000 5 4 K 3 7500 5 425
4 0 ð1 K 3 Þ 7500 2

If we have the costs as described in Table 1, the perturbation in cooling with retarded actions, as described
above, gives the final revenue for good PP (ordinary price and sales), reduced for the costs of shortage, as
presented in Table 2.

5. Conclusion

The cold chains management requires very careful temperature control and quick reactions when
perturbations appear in temperature or time delays occur. In our example it is presented how sensitive such
a chain can be. Visibility and proper control is needed. We are able to conclude that the analysis in the
frequency space can be straight forward, more than the analysis in the time domain. Using input–output
analysis, Laplace transforms and MRP approach, the problem can be transferred to LP parametric
problem which enables a clear presentation of the results of actions against perturbations, that can appear
on the different stages of a cold chain. The same approach, which was presented here for evaluation of the
retarded changes in temperature, can be used also in the evaluation of the actions against the perturbations
of lead times.

References

Baldiwala, Q., 2001. Developing a global supply chain. Logistics Spectrum 35/4, 25–29.
Bogataj, L., Bogataj, M., 2001. Minimization of distribution costs and response time in global supply chain defined in the frequency