Professional Documents
Culture Documents
Abstract
In the food-processing industry, usually a limited number of storage tanks for intermediate storage is available, which
are used for different products. The market sometimes requires extremely short lead times for some products, leading to
prioritization of these products, partly through the dedication of a storage tank. This type of situation has hardly been
investigated, although planners struggle with it in practice. This paper aims at investigating the fundamental effect of
prioritization and dedicated storage in a two-stage production system, for various product mixes. We show the
performance improvements for the prioritized product, as well as the negative effects for the other products. We also show
how the effect decreases with more storage tanks, and increases with more products.
r 2007 Elsevier B.V. All rights reserved.
Keywords: Food processing; Two-stage production; Intermediate storage; Product prioritization; Dedicated storage
0925-5273/$ - see front matter r 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpe.2006.12.018
ARTICLE IN PRESS
44 R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353
orders for one customer. Moreover, tanks are that get prioritized to be delivered within a relatively
usually limited in number and size, as high short lead time.
investments are involved for this type of storage The overall contribution is to better understand
facilities. The time of storing an unpacked product the intermediate storage in typical food-processing
is limited by its shelf life. companies in order to improve planning and
A main complication is, however, that usually the scheduling in such situations and to improve
number of products exceeds the number of tanks. decision making with respect to the required
Storing a product is thus more than just allocating a number of tanks. In general, the situation with
production batch to an arbitrary tank. On the one intermediate storage can be assessed using a
hand, availability of products for packaging is common performance measure like lead time. There
needed, leading to the wish to ll as many tanks are two specic effects of interest: blocking and
as possible with basic product. On the other hand, starvation. Blocking refers to the non-availability of
availability of empty tanks is required to enable storage tanks for nished product which has to wait
continuous processing in the rst stage of the in the processing stage, while starvation means idle
production process. Planners tend to believe that capacity in the packaging stage due to non-
building extra tanks is the solution for this problem, availability of basic product. For instance, in the
but, as said before, that is expensive. What makes situation described in this paper, blocking happens
this situation even more complex is the fact that if a batch is produced in the rst stage, but no
market demands can be different among products. intermediate storage tank is available for the
Lead times for products can be under extreme product. Then the product has to remain in the
pressure, which creates a situation where certain batch processor, which delays further batch proces-
products need to get priority over other products. sing until a storage tank becomes available. Possible
This prioritization often results in xed assign- prioritization and storage dedication have a large
mentsor dedicationof storage to the prioritized impact on these blocking and starvation effects,
product. In this paper, we specically look at the which in turn highly inuence the behaviour of a
effects of allocation policies for storing products in production system with limited intermediate sto-
tanks, based on product prioritization. The litera- rage.
ture in operations management hardly pays atten- The remainder of this paper is organized as
tion to this important decision area. follows. Section 2 gives some background informa-
The aim of this paper is to address the effect of tion on previous research. In Section 3, the
prioritization of a product versus treating all production model studied in this paper is described.
products equally. An important result of the Subsequently, a deterministic analysis of the pro-
prioritization is a specic type of storage allocation: duction system is presented in Section 4. Following
the permanent allocation, further addressed as that, Section 5 presents a numerical study and its
dedication, of a tank to a prioritized product. This results. Finally, Section 6 presents conclusions and
type of storage allocation can also be found in suggestions for further research.
situations where production is hybrid make-to-
order (MTO) and make-to-stock (MTS), which is 2. Background
quite common in the food-processing industry
(Soman et al., 2004). In those situations, the In the food-processing industry, reducing lead
decision to make a product to stock or to storage times is becoming increasingly important as im-
is mainly based on its share in the product mix; proved customer service is important, especially
high-volume products are normally MTS, and low- when dealing with powerful food retail chains (see,
volume products MTO (see, e.g., Youssef et al., e.g., Meulenberg and Viaene, 1998). Das and Abdel-
2004). However, this decision can also be forced on Malek (2003) also investigate the effects of a varying
the company by market demands. Therefore, we lead time in a supply chain on exible delivery. They
specically investigate the effects of prioritization state that lead times are one of the main causes for
by means of dedication policies for various shares of supplierbuyer grievances in a supply chain. As
a product in the product mix. such, reducing lead times creates more pressure on
With the present study we are able to assess the these relationships in the supply chain.
overall effect on systems performance of dedicating Lead time reduction also relates to the current
a tank for low-demand and high-demand products interest in hybrid MTOMTS production systems
ARTICLE IN PRESS
R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353 45
(see, e.g., Huiskonen et al., 2003; Soman et al., explicitly discussed, the literature pays some atten-
2004). For the food-processing industry, a signi- tion to the issues of dedication and exibility in a
cant share of the production is customer-specic, qualitative sense.
which often results in a large MTO part in their The main objection against dedication of storage
production system. The reducing lead times inter- tanks might be the loss of exibility. One might
fere with these policies, as it is no longer possible to assume that without dedicated storage, assigning
produce the required product from raw materials products to tanks is easier and results in higher
within this lead time. The answer usually lies in the performance of the overall production system. If
storage of certain basic products, which can be each product has its own tank, assignment is even
packaged for customer-specic orders. This results easier. However, in food processing, the number of
in a hybrid MTOMTS system at the intermediate products usually exceeds the number of storage
storage. tanks, so only a partial dedication is possible. In the
In the literature on hybrid MTOMTS systems literature, dedication has hardly been discussed, but
(see Soman et al., 2004, for an overview) demand exibility (as being its natural opposite) has been
characteristics (e.g., the share of the product in the extensively treated. The main question seems to be
product mix) are mostly used to determine whether how much exibility should be added, as it is
products should be made to order or to stock. As assumed that exibility and exible equipment are
Soman et al. (2004) also argue, other market more expensive. For instance, Jordan and Graves
characteristics are often ignored. In our study, we (1995) develop principles on the benets of process
focus on one specic characteristic: lead time. A exibility. One of the main outcomes is that a small
short lead time requires MTS at the intermediate amount of exibility can have almost the same
storage level and prioritization of the product to be benets as total exibility. In other words, after a
able to meet the required lead time. This is closely certain exibility is reached, there are rapidly
related to the work of Sox et al. (1997), who denote decreasing benets when adding additional exibil-
this required lead time with their service window. ity. This argument might be transformed for
Sox et al. (1997) then prioritize the MTO products dedication of storage tanks by posing that removing
to ensure a good overall customer service. They also some exibility could initially be relatively harmless
note that when the service window becomes very to production performance. However, this is less
short (compared to the average ow time of the likely in situations where only a small number of
factory), prioritization degrades performance. In storage tanks are available, as the dedication of one
our study, the reason for prioritization and dedica- of those tanks removes a signicant amount of
tion is the fact that the required lead time (or service exibility.
window) is shorter than the average ow time of the In summary, the above discussion clearly shows
factory. Therefore, we explicitly aim at investigating that the effects of prioritization and the decision to
the effect of prioritization and dedication (and dedicate storage have not yet been systematically
treating all products equally with exible storage investigated. For production planning and schedul-
allocation as the alternative policy) on the perfor- ingand also for the (re)design of production
mance of a production process. processesit is important to understand these
Next to the prioritization of a product, dedicated effects.
storage in the intermediate storage facility is also
required to meet the demand. In the literature, we 3. Production model
see that several papers address intermediate storage
in scheduling. Most papers develop techniques to 3.1. Production system
incorporate these storage tanks in mathematical,
mostly MILP-based, scheduling models (e.g., Belarbi The production system studied in this paper
and Hindi, 1992; Ha et al., 2000; Rajaram et al., consists of two distinct stages, connected by
1999; Yi et al., 2000). In the majority of these intermediate storage tanks (see Fig. 1). The rst
papers, the distinction between dedicated and production stage concerns a (non-preemptive) batch
exible storage is mentioned and considered in the process with a single batch processor with a xed
techniques developed. However, this distinction is batch size B. This xed batch size resembles a
assumed to be predetermined and known. While the technical constraint that is often encountered in the
decision to dedicate a storage tank or not is not food-processing industry (e.g., kettle size). Due to
ARTICLE IN PRESS
46 R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353
Basic
food product
Tanks End products
Raw material Batch Product
processing packaging
Fig. 1. General form of the two-stage production process with intermediate storage tanks in the food-processing industry.
variability in raw material quality, processing times Packaging can only start if a customer order has
are variable (see Fransoo and Rutten, 1994). arrived: packaging is customer-specic and or-
In the second stage, the intermediate food der-specic.
product is packaged in small and large packaging
sizes, depending on customer orders. This translates
3.2. Production scheduling
into different packaging times; for small sizes more
time is required to package a certain amount of
Customer orders arrive continuously during the
product (e.g., more packages moving through the
day. They have several distinct characteristics: (i)
line). This variability inuences performance
product type, (ii) packaging size, and (iii) arrival
through the blocking and starvation effects men-
time.
tioned in the Introduction.
To study the effect of dedication, we use two
The intermediate storage consists of K storage
different storage policies:
tanks, which are used to store N different inter- 8
mediates. An important aspect is that in a storage >
> F a fully flexible policy; in which every tank can
>
>
tank only one production batch can be stored < be used for every product;
concurrentlyeven in case of the same product. P
>
> D a policy in which one storage tank is dedicated
>
>
This is due to (i) traceability requirements and (ii) not : to a specific prioritized product:
mixing batches to ensure quality. This also makes the
size of the storage tanks irrelevant, as long as they (1)
can at least contain one batch of product from the Without loss of generality, product 1 can be used as
processing stage. We assume this is the case. the prioritized product with a dedicated storage
Furthermore, we make the following assump- tank in policy D (also referred to as the dedicated
tions: product).
For policy F , the arriving orders are collected in
The production system operates in one daily shift an orderpool until a full batch of a certain product
of 8 h. can be produced in the rst stage. The batch order
Raw materials for the batch processing stage, as is then placed in an FCFS (rst-come-rst-serve)
well as packaging materials for the packaging queue for the batch processor while the orders are
stage are always available with negligible lead moved from the orderpool to the queue at the
times. packaging line.
Products immediately leave the production sys- For policy D, a runout time procedure is used for
tem after packaging. the dedicated product, because we need to keep this
Transportation time to and from storage tanks is product on stock on the intermediate storage level.
negligible. The batch order for the dedicated product (a
No changeover times for processing, packaging, replenishment order) is generated when the runout
and storage. time of the content of the dedicated tank is smaller
The quality of the product remains constant in than the average batch processing time. The runout
the storage tanks for a xed period, after which it time is calculated as follows for product i:
is discarded at no extra cost (see also Nahmias,
ROi I i Oi =Di , (2)
1982; Raafat, 1991, for discussions on modeling
perishability). where I i is the inventory level of product i in the
Dedication of a storage tank is assumed to be intermediate storage, Oi is the number of waiting
implemented to assure a short lead time for the orders for product i, and Di the average number of
prioritized or dedicated product. orders arriving per time unit.
ARTICLE IN PRESS
R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353 47
Because orders for the dedicated product (in policy 4.1. Dedication for a product with a share of 33%
D) are immediately packaged from intermediate
storage, arriving orders for this product move straight The rst scenario we analyse is that of equal
to the packaging queue. For the other (non- demand for all products.
dedicated) products, the orders are processed like In Fig. 2, two excerpts from Gantt charts
in policy F (collected until a full batch is realized). In illustrate the system behaviour. For policy F , we
the batch order queue, the product with dedicated see that a cyclic production pattern emerges, which
storage has priority over the other products (to ensure only needs one of the two storage tanks. The second
the timely replenishment and short lead time). Gantt chart in Fig. 2 shows that this situation
For the second stage, a basic sequencing rule is changes dramatically when policy D is implemented.
used for scheduling the packaging line. The Several important aspects in this chart are (i) the
customer order in the packaging queue with the possibility to package orders for product 1 from
earliest arrival time is packaged rst (FCFS). If intermediate storage; (ii) the occurrence of blocking
the required basic product is not (yet) available, the at the batch processor due to unavailable exible
next order in the queue is selected. In case of policy storage; and (iii) the fact that two storage tanks is
D, this FCFS rule comes second after a priority rule getting restrictive, while only one was needed in the
for the dedicated product. exible case. This results in an unbalanced situation,
characterized by an increasing backlog of orders in
the long run.
4. Deterministic analysis
To explore the described system, we will perform 4.2. Dedication for a product with a share of 10%
a deterministic analysis of the behaviour of the
simplest system conguration (K 2 storage tanks Here, we assume that a storage tank is dedicated
and N 3 basic products) that still enables us to to a product that only represents a small fraction of
study the effects of dedication for several scenarios the product mix. Reduced lead times in the supply
with different product shares for the dedicated chain might be the main reason.
product. Two storage tanks are needed to be able to Fig. 3 shows partial schedules for policies F and
distinguish between dedicated and exible storage; D for a situation where product 1 covers 10% of the
three products are the minimum to have more than product mix, and products 2 and 3 together cover
one product in the exible storage. For the sake of the additional 90%. For policy F , the schedule is
simplicity, all possible variability (in order arrivals, still cyclic in nature, albeit that the cycle is getting
processing times, packaging times) is ignored and rather large. In principle, products 2 and 3 are
we assume a utilization of 100%, which we achieve alternating, with one batch of product 1 being
by setting the order arrival rate equal to the produced every 10 batches. For policy D, we see
production capacity. that indeed the demand for product 1 can be met in
Time
BatchProc
Tank 1
F
Tank 2
PackLine
Time
BatchProc
Tank 1
D
Tank 2
PackLine
Legend:
Product 1 Product 2 Product 3 Blocking
Fig. 2. Gantt charts that illustrate the effect of dedication of intermediate storage for a product with a share of 33%.
ARTICLE IN PRESS
48 R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353
Time
BatchProc
Tank 1
F
Tank 2
PackLine
Time
BatchProc
Tank 1
D
Tank 2
PackLine
Fig. 3. Gantt charts that illustrate the effect of dedication of intermediate storage for a product with a share of 10%.
Time
BatchProc
Tank 1
F
Tank 2
PackLine
Time
BatchProc
Tank 1
D
Tank 2
PackLine
Fig. 4. Gantt charts that illustrate the effect of dedication of intermediate storage for a product with a share of 50%.
a package-to-order fashion. However, this again 4.4. Dedication for a product with a share of 60%
results in blocking effects at the batch processor,
and signicant starvation effects at the time the Finally, we study a situation in which the rst
dedicated tank needs to be relled. This is again an product has the largest part of the demand: 60%.
unbalanced situation, in which the high utilization The other products are each at 20%. Corresponding
rate creates an ever-increasing backlog. partial schedules are shown in Fig. 5.
The cyclic schedule still remains for policy F (here
4.3. Dedication for a product with a share of 50% it is 11123). For policy D, the system is again
getting unbalanced through blocking and starvation
Here we assume that a high-demand product is effects. This makes a growing backlog of orders.
stored in a dedicated tank. The reason could simply
be the convenience in scheduling if a certain product 4.5. Concluding remarks
always goes to a specic tank. This product has
50% of the demand, while products 2 and 3 each The above analysis yields a better understanding
have 25%. Fig. 4 shows excerpts from the corre- of how exible and dedicated assignment of storage
sponding schedules. inuences systems performance. It shows that
As in the previous scenarios, policy F results in a exible assignment of products to tanks (policy F )
cyclic schedule (now 1213), which only utilizes results in cyclic schedules, while policy D creates
one storage tank. Policy D, however, shows more irregular schedules. While lead times for the
different results. Here, the package-to-order possi- dedicated product are smaller it negatively affects
bility is again visible, but no blocking occurs at the overall performance and lead times of the other
batch processor. Both storage tanks are highly products. For some scenarios (50% product 1), the
utilized, and the production system is in balance. results showed balanced production systems for
ARTICLE IN PRESS
R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353 49
Time
BatchProc
Tank 1
F
Tank 2
PackLine
Time
BatchProc
Tank 1
D
Tank 2
PackLine
Fig. 5. Gantt charts that illustrate the effect of dedication of intermediate storage for a product with a share of 60%.
policy D. For other scenarios (33% product 1, 10% varied from 10% to 90%. The remaining products
product 1, 60% product 1), the production system all have an equal part of the remaining share. This is
got unbalanced due to blocking and starvation calculated as follows:
effects, which in the long term results in an 100 S 1
increasing backlog of orders. Si 8i 2; . . . ; N. (3)
N 1
5. Numerical experiments In the experiments, this is modeled by using the
product mix shares as probabilities for the arriving
To account for variability in processing times, orders.
packaging times, and order arrivals, this section will The main performance criterion used in this
present numerical experiments to further analyse paper is average lead time, which is calculated as
the differences between policies F and D. We study the time in minutes between the arrival of an order
several system congurations, which should provide and the completion of that order in the packaging
insight into the interaction between the product mix stage. These average lead times are calculated for
and the system performance for both policies. each of the products, to investigate the effects of
different product mixes. Next to lead time, blocking
5.1. Experimental design will also be used as one of the important underlying
aspects of lead times.
The experimental factors to be varied are (i) the The simulations are performed in MATLAB. Before
number of storage tanks, (ii) the number of basic the experiments are conducted, a warmup period of
products, (iii) the dedication policy, and (iv) the one month is determined using a graphical method
product mix. with average lead times. Furthermore, run lengths
The number of storage tanks (K) is varied from 1 of one year with ve replications are determined
to 10 and the number of basic products (N) from 2 using a 95% condence interval for the average lead
to 10. We expect that the effect of dedication is time. This results in relative half-widths of the
stronger if there are more products than storage condence interval between 0.8% and 5%. Before
tanks, but we also investigated other scenarios. In the actual simulation runs are conducted, numerous
the paper, we present only scenarios where N4K, test runs for different parameter settings are
as these are the situations where the dedication and performed, while closely watching the systems
prioritization has the biggest effects. Furthermore, behaviour for verication purposes.
in practice, the number of products is normally
larger than the number of intermediate storage 5.2. Parameter settings
tanks.
The two different dedication policies were already Customers orders arrive continuously according
presented in Eq. (1) in Section 4. For the product to an exponential distribution with l 0:16 (per
mix, nine different situations will be considered. The minute). This resembles a Poisson process with
share of product 1 (S 1 ) in the product mix will be certain interarrival times (6.25 min), which is a
ARTICLE IN PRESS
50 R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353
5.3.1. Basic configuration expect this follows from the regular arrival of batch
For policy F , there is a difference in the lead time orders in this symmetric situation. A more irregular
between product 1 and the other products for small arrival process is likely to create more variety in the
and large values of S 1 (see Fig. 6). This is the result length of the batch queue. This even distribution of
of the difference in waiting time before a batch can the product mix possibly results in an efcient cyclic
be formed. For example, if S 1 is very high, it takes production schedule, as would be used in practice in
less time to collect enough orders to form a batch of similar situations, and is also reected in the
product 1. This results in smaller lead times. As seen deterministic analysis.
in the deterministic analysis, a batch of product 1 The results for policy D show rather different
will occur more often in the production cycle. curves (see Fig. 7). Firstly, as expected, a signicant
Another aspect is that the more the asymmetry in decrease of the lead time for product 1 is achieved
the product mix, the higher are the lead times. (compared to policy F ), and the higher the share S 1 ,
Partly, this is related to the time until a batch is the lower the lead time due to decreasing runout
formed. However, analysis of the amount of times and less interference with other products. For
blocking shows that for values of 3040% for S 1 , these other products, policy D results in an overall
the lowest amount of blocking is experienced. We increase in lead times. Especially when these
ARTICLE IN PRESS
R. Akkerman, D.P. van Donk / Int. J. Production Economics 108 (2007) 4353 51
Sox, C.R., Thomas, L.J., McClain, J.O., 1997. Coordinating International Journal of Production Economics 69 (3),
production and inventory to improve service. Management 297306.
Science 43 (9), 11891197. Yi, G., Suh, K.-H., Lee, B., Lee, E.S., 2000. Optimal operation of
Van Dam, J.P., Gaalman, G.J.C., Sierksma, G., 1998. Designing quality controlled product storage. Computers & Chemical
scheduling systems for packaging in process industries: a Engineering 24 (27), 475480.
tobacco company case. International Journal of Production Youssef, K.H., van Delft, C., Dallery, Y., 2004. Efcient
Economics 5657, 649659. scheduling rules in a combined make-to-stock and make-to-
Van Donk, D.P., 2001. Make to stock or make to order: order manufacturing system. Annals of Operations Research
the decoupling point in the food processing industries. 126 (14), 103134.