Predictive Analytics for Supply Chains: a Systematic
Literature Review
Kevin Bonnes
University of Twente
P.O. Box 217, 7500AE Enschede
The Netherlands
K.G.Bonnes@student.utwente.nl
ABSTRACT
Predictive analytics is a group of methods that uses sta-
tistical and other empirical techniques to predict future
events, based on past occurrences. Predictive analytics
can generate valuable information for the management
of a supply chain company to improve decision-making.
Even though the importance of the topic is clear, there is
no clear overview of the use of predictive analytics in the
supply chain currently. This research provides a state-of-
the-art by performing a systematic literature review. In
the literature review we have found the models, methods,
techniques and applications of predictive analytics in the
supply chain and have determined the trends and litera-
ture gaps. The most important finding is that even though
there is only a limited amount of literature available, the
interest in this topic is growing gradually. We also pro-
vide future research directions for further research on this
subject.
Keywords
Literature Review, Predictive Analytics, Prediction, Sup-
ply Chain, Big Data, Forecasting
1. INTRODUCTION
Predictive analytics is a group of methods that uses sta-
tistical and other empirical techniques to predict future
events, based on past occurrences. Although predictive
analytics has become a hot topic recently, there are many
companies within the supply chain sector that dont make
use of it yet [61]. As literature states, enhancing effective-
ness and efficiency of supply chain analytics is a critical
component of a companys ability to achieve competitive
advantage [61]. Predictive analytics for supply chains is
therefore an interesting topic.
Predictive analytics are used in many different fields. This
research focuses on the supply chain sector. An example
of the use of predictive analytics in the supply chain is
predicting the arrival times of a truck or a ship. There
are multiple variables that can influence arrival times, for
example: the weather, traffic congestion and the drivers
driving style. If the reliability of predicting arrival times
can be increased, costs can be saved within the entire
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies
are not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. To copy oth-
erwise, or republish, to post on servers or to redistribute to lists, requires
prior specific permission and/or a fee.
21st Twente Student Conference on IT June 23rd , 2014, Enschede, The
Netherlands.
Copyright 2014, University of Twente, Faculty of Electrical Engineer-
ing, Mathematics and Computer Science.
1
supply chain. These costs can be found in waiting times
for trucks in distribution centers and improved decision-
making.
If companies get a better understanding of how predictive
analytics should be used, decision-making on management
level will be improved. Currently there is literature avail-
able on this research area but there is no overview. This
research will provide this overview by mapping out all of
the relevant information available by conducting a system-
atic literature review, as described by Webster & Watson
[78]. Afterwards, the literature will be categorized and
presented in terms of models, methods, techniques and
applications. We will also show the current trends in pre-
dictive analytics for the supply chain, to see where the
field is going.
1.1 Problem statement
Even though there is a large body of work available on
both predictive analytics and the supply chain, there has
not been an overview of the combination of these subjects.
This literature study will fill that gap by providing insight
on the current use of predictive analytics for supply chains.
1.2 Research questions
Based on the problem statement, we arrive at the following
research questions:
RQ1: What are the models, methods, techniques and
applications used in predictive analytics for the supply
chain?
RQ2: What are the research gaps of the literature
found?
1.3 Remainder of this paper
The remainder of this paper is organized as follows: Sec-
tion 2 describes the research method. In section 3 the
results are presented. Section 4 provides a discussion of
the results, the current trends and the gaps in the litera-
ture. Section 5 presents the conclusions, limitations and
provides future research directions.
2. RESEARCH METHOD
In order to answer the research questions and validate the
findings a systematic literature review was performed, fol-
lowing the method described by Webster & Watson [78].
This method was chosen to increase the scientific value of
the paper by making sure that all of the relevant litera-
ture on this subject is taken into account. In this section
each step of the literature search will be described and mo-
tivated so that our findings can easily be replicated and
validated. In order to keep the quality of the research as
high as possible our literature search was limited to only
those databases, journals and conferences that have a good
academic reputation.
Literature was searched in the databases Scopus and Web terms of concepts. Categorizing the literature plays an
of Science. These are two of the leading databases and important role as it gives a good overview of the available
provide good coverage of the relevant sources. literature and the literature gaps. Categorization is done
The next step in the process was carefully selecting the on four different aspects of the literature found: models,
journals and conferences, to make sure that we only in- methods, techniques and applications. These four aspects
clude the relevant literature and dont exclude any rele- have been chosen because we think they are the most valu-
vant paper by omitting a quality publication. Relevant able as they give a good understanding of why and how
journals were determined by creating a list of relevant re- predictive analytics should be used.
search areas and selecting the top journals in each of these
research areas. A list of the research areas can be found 3.1 Models
in Appendix A. Models in predictive analytics for supply chains can be di-
vided into two categories: explanatory models and predic-
For each of the research areas a list of journals was de- tive models [66, 67]. Explanatory models (often statistical
termined based on the top 20 journals ranked by impact models) test a causal hypothesis based on historical data,
factor, from the ISI Journal Citation Ranking [72]. Irrele- and provide a conclusion about the model. They do not
vant journals (based on topic) and duplicate journals were use new data to make predictions. Predictive models are
excluded. The list of journals was verified and validated also based on historical data, but make use of new data
by several existing journal rankings [4, 23, 46, 47, 51, 63, to test the model and make predictions about the future.
64, 69, 71]. For each of these rankings the top 20 was com- Of models found in the 64 papers reviewed, only 18 were
pared with the ISI JCR top 20 and missing journals were categorized as predictive models. Even though explana-
added. This was done to make sure that we did not miss tory models do not imply predictive power [67], we do
out on any important journal in any of the research areas. take them into account mainly because of a lack of real
Eventually, a list of 137 journals was created. The journal predictive models. An overview can be found in Table 1.
list and the validation can be found in Appendix B.
Conference papers are generally of less importance com-
pared to journals in the business/SCM and information Table 1. Models
systems disciplinaries, except for the computer science re- Type of Amount References
search area [73]. Conference papers however have a shorter model of papers
throughput time and are thus often more recent and pro- Explanatory 46x (72%) [3, 610, 1214, 16, 18
vide a better state-of-the-art than journal papers [28]. model 22, 2427, 30, 32, 37, 39
Hence, a total of 51 important conferences for the com- 43, 45, 5254, 56, 58
puter science and information systems area were also re- 60, 62, 65, 70, 76, 77, 7984]
viewed. A list of the selected conferences can be found Predictive 18x (28%) [1, 2, 5, 11, 15, 17, 29, 31,
in Appendix C. This list was based on several conference model 34, 35, 38, 44, 49, 50, 55,
rankings by field rating [33, 36, 48]. To ensure the nov- 57, 68, 75]
elty of the conference papers found we specified an extra
criterion: papers have to be published in 2010 or later.
3.2 Methods
Based on the research questions a research query was for- The methods used in the literature are made up of two
mulated. The research query was formulated in several parts: the data collection methods, either empirical or
steps of improvement, to make sure we did not miss or simulation, and the analysis techniques which will be dis-
exclude any relevant papers. The research query can be cussed in the next section.
found in Appendix D. To increase the likeliness of the re-
sults being relevant, the keywords in the research query Data collection methods are the methods used to collect
have to occur in the titles, abstracts or keywords of the data as an input for the model. We categorized the data
articles. collection method in two different types: methods using
empirical, measured data and methods using generated
With the research query and the journal- and conference- data. An overview of the data collection methods can be
query, a literature search was performed. In Scopus, a found in Table 2.
total of 446 articles were found. In Web of Science, we
found a total of 429 articles. After removing the duplicate
articles that were found in both Scopus and WoS, a total Table 2. Data collection methods
of 581 papers remained.
Category Methods Amount References
3. RESULTS of pa-
The results of the literature search were evaluated on ab- pers
stracts by two reviewers, to make sure that the papers that Measured Empirical 53x [3, 5
had no relation with both the supply chain and predictive data / Industry (83%) 15, 17, 18, 20
analytics were excluded. After the selection process a to- data, Exper- 22, 26, 27, 29
tal of 160 papers remained. The remaining set of papers iment, Case 32, 34, 35, 37
was read and evaluated more in depth, until a final set of study, Survey 45, 49, 50, 52,
64 papers was composed. This was done on the following 57, 59, 60, 62,
criterion: Does the paper formulate a clear model that can 65, 68, 70, 75
be used for predictive analytics in the supply chain sector? 77, 7984]
Generated Simulation 11x [1, 2, 16, 19,
Based on the final collection of papers, a concept-centric data (17%) 24, 25, 53
literature review was performed. A concept-centric liter- 56, 58]
ature review is a literature review in which the concepts
determine the structure of the review rather than the au-
thors [78]. This means the results will be categorized in

2
3.3 Techniques
Table 4. Techniques vs. Models
Analysis techniques (e.g. time series, regression), also of-
Technique / Explanatory Predictive Total
ten referred to as research methods, are the underlying
Model model model
techniques used to analyze the data. Models in predictive
Time series 13x (93%) 1x (7%) 14x
analytics are always based on one or more techniques. An
(100%)
overview of the techniques used in the literature can be
Regression 6x (86%) 1x (14%) 7x
found in Table 3. In Table 4, the top 5 techniques are
(100%)
shown with reference to the models.
Neural net- 4x (67%) 2x (33%) 6x
works (100%)
Mixed inte- 2x (40%) 3x (60%) 5x
ger program- (100%)
ming
Exponential 5x (100%) 0x (0%) 5x
smoothing (100%)

Table 3. Techniques
Technique Amount of pa- References
pers
Time series 14x (19%) [1, 14, 1921, 45, 3.4 Applications
54, 56, 60, 76, The applications of predictive analytics for supply chains
77, 79, 81, 83] are also categorized. These can be found in Table 5. This
Regression 7x (9%) [5, 7, 16, 18, 22, table gives an overview of the usage of predictive analytics
59, 65] within the supply chain.
Neural networks 6x (8%) [20, 31, 43, 50, As shown in Table 6, there are three major applications of
80, 82] predictive analytics within the supply chain: forecasting,
Mixed integer 5x (7%) [2, 6, 17, 34, 40] inter-organizational collaboration and monitoring. In Ta-
programming ble 6 for each of these applications the amount of papers
Exponential 5x (7%) [19, 60, 77, 79, found is shown with reference to the two types of models.
smoothing 83]
Fuzzy logic 4x (5%) [24, 27, 30, 42]
Simulation 4x (5%) [3, 12, 70, 75]
Statistical tech- 4x (5%) [15, 29, 35, 52]
niques
Mathematical 3x (4%) [25, 26, 39]
techniques
Structural equa- 3x (4%) [8, 9, 57]
tion modelling
Support vector 2x (3%) [6, 45] Table 6. Applications vs. Models
machines Application / Explanatory Predictive Total
System dynam- 2x (3%) [53, 58] Model model model
ics Forecasting 27x (79%) 7x (21%) 34x
Partial least 2x (3%) [9, 44] (100%)
squares Inter- 10x (77%) 3x (23%) 13x
Logistic regres- 2x (3%) [32, 49] organizational (100%)
sion collaboration
Linear regres- 2x (3%) [13, 40] Monitoring 6x (55%) 5x (45%) 11x
sion (100%)
Markov predic- 2x (3%) [37, 38]
tion
Markov decision 1x (1%) [11]
processes
Markov chain 1x (1%) [10]
method
Nonparametric 1x (1%) [68]
regression
Continuum ap- 1x (1%) [55] 3.5 Trends
proximation Besides the tables presented above, 4 line charts were gen-
Analytic hierar- 1x (1%) [41] erated to be able to detect trends. The line charts show
chy process the amount of relevant papers for each year, for both ex-
Rough set the- 1x (1%) [84] planatory models and predictive models. Figure 1 presents
ory approach a general chart and Figure 2, 3 and 4 present line charts
Machine learn- 1x (1%) [62] respectively for each of the applications forecasting, inter-
ing organizational collaboration and monitoring.

3
 Table 5. Applications
Category Topic(s) Amount of papers References
Forecasting Demand/order forecast- 34x (51%) [1, 3, 9, 1214, 17, 19
ing, inventory forecasting, 21, 29, 31, 35, 37, 41, 43,
predicting bullwhip effect, 45, 49, 5254, 56, 58
scheduling, predicting 60, 62, 70, 7577, 79, 8183]
production volatility
Inter-organizational collab- Predicting influence of 13x (19%) [5, 8, 13, 15, 18, 22, 27, 30,
oration collaboration, supplier 32, 42, 57, 65, 84]
selection, predicting sup-
ply chain integration,
predicting stakeholder
collaboration strategies
Monitoring Container tracking, travel 11x (16%) [6, 11, 16, 25, 26, 34, 39,
time prediction, logistics 44, 55, 58, 68]
planning, predicting order
arrival, predicting flow
times, predicting RFID
performance, inventory
tracking, predicting logistic
costs
Price forecasting Price forecasting 2x (3%) [13, 38]
Product deficiency Error prediction, predicting 2x (3%) [2, 10]
product failure
Other Predicting supply chain 5x (7%) [7, 24, 40, 50, 80]
performance, predicting
cost-to-serve for new cus-
tomers, predicting price
negotiation outcomes, pre-
dicting customer purchases,
predicting critical factors in
a supply chain

10 10
explanatory explanatory
trend trend
8 predictive 8 predictive
trend trend
6 6
Papers

Papers

4 4

2 2

0 0
1995 2000 2005 2010 2015 1995 2000 2005 2010 2015
Year Year

Figure 1. Models vs. Time Figure 2. Models vs. Time, Forecasting

4
 10 measured data. As expected the percentage of papers us-
explanatory ing simulated data as input is quite low compared to the
trend percentage of papers using measured data.
8 predictive
trend 4.3 Techniques
As shown in Tale 3, predictive analytics consists out of
6 multiple different techniques. Some models use a combi-
Papers

nation of techniques (e.g. [6, 20, 60]). By far the most

used technique is time series. Time series techniques are
4 used to identify a pattern in a sequence of data points.
Other widely used techniques are regression, neural net-
works, exponential smoothing (a technique used in com-
2 bination with time series) and Mixed integer (non-)linear
programming.
In Table 4, the techniques are shown with reference to the
0 two types of models. As shown, the most used technique,
1995 2000 2005 2010 2015
time series, is almost always used in explanatory models.
Year The percentage of papers using time series techniques in
a predictive model is only 7%. This number is quite low
Figure 3. Models vs. Time, Inter-organizational compared to the average of 28% of predictive models in
collaboration general (Table 1). Besides time series techniques, this also
applies to regression techniques and exponential smooth-
ing techniques. The only technique within the top 5 that
10 is being used in an exceptionally high percentage of pre-
explanatory dictive models compared to explanatory models is mixed
integer programming. In general, the percentage of predic-
trend
8 tive models compared to explanatory models for the top 5
predictive
techniques is only 19%, compared to the 28% of Table 1.
trend
6 4.4 Applications
Papers

For applications in predictive analytics for the supply chain

4 zational collaboration and monitoring. 86% of the papers
2 ics such as demand forecasting, inventory forecasting and
0 activity in which a company tries to predict the quantity of
1995 2000 2005 2010 2015
Year
Figure 4. Models vs. Time, Monitoring
4. DISCUSSION
4.1 Models
As shown in Table 1, out of the 64 papers only 18 of them
included a predictive model. This means only 18 papers
actually used new data to make predictions about the fu-
ture based on the model. The other papers presented ex-
planatory models which can be used to make predictions
but did not directly apply the model to new data. This
means the 46 papers that presented an explanatory model
did not directly make use of predictive analytics, but in-
stead only presented a model, which implies that true an-
alytics are scarce in the supply chain field.
4.2 Methods
Data collection methods in predictive analytics are mainly
focused on empirical, measured data. As shown in Table
2, 53 of the 64 papers make use of measured data to test
their model. Measured data consists out of empirical or
industry data, often generated by methods such as ex-
periments, case studies or surveys, or a combination of
these methods. Simulated data is data generated by sim-
ulations. Simulated data is often used due to a lack of
there are three major categories: forecasting, inter-organi-
can be classified in one of these three categories.
The first category, forecasting, encompasses several top-
predicting the bullwhip effect.
Demand forecasting (e.g. [14, 43, 83]) is the forecasting
products the customer will purchase. Predictive analytics
are used to predict future demands based on historical
demand information. Demand forecasting is often done
by applying time series to a given set of input data.
Inventory forecasting is closely related to demand forecast-
ing. With inventory forecasting a company tries to pre-
dict the amount of inventory required to fulfill customer
orders. Examples of papers about inventory forecasting
are [59, 77].
Predicting the bullwhip effect is also related to demand
forecasting. The bullwhip effect is the problem that arises
due to a continuously changing demand of orders, which
makes it hard to predict the amount of produced prod-
ucts. Examples of papers presenting models to predict
the occurrence of the bullwhip effect are [53, 54, 56].
The second category is the inter-organizational collabo-
ration applications. Examples of inter-organizational col-
laboration applications are supplier selection and predict-
ing the influence of collaboration on supply chain perfor-
mance. Supplier selection is about predicting the best pos-
sible supplier based on historical information about similar
situations. Examples of papers providing a model for op-
timal supplier selection are [27, 30, 42]. Predicting the
influence of supply chain collaboration is about sharing
information between supply chain partners to improve the
overall performance of the supply chain. Examples of pa-
pers are [5, 8].

5
The third category is monitoring applications. Examples
of monitoring applications are travel time prediction and
order arrival time prediction. Both of these applications
are closely related. Travel time prediction is often used
to predict order arrival times. Predictive analytics in this
application are used to predict travel times based on his-
torical information of similar routes, taking into account
the influence of external factors such as the weather and
traffic congestion. Examples of papers about travel time
prediction or arrival time prediction are [11, 68]. Another
example of monitoring is predicting transport costs. Ex-
amples of papers are [16, 55].
As shown in Table 5, a lot of research is done in the fore-
casting application compared to the other applications.
However, most of the research done in the forecasting ap-
plication only provides an explanatory model, as Table 6
shows. There is a limited amount of research that actually
provides predictions. On the other hand, in the monitor-
ing application field the amount of predictive models is
relatively high.
4.5 Trends
Another aspect of our literature study is determining the
trends of predictive analytics for supply chains. Figures
1 to 4 show the amount of papers published per year per
application, together with their trend lines. There are a
number of conclusions that can be drawn from these four
figures.
First of all, as shown in Figures 1 to 4, predictive analytics
for the supply chain is still a relatively premature research
area. There is a limited amount of research available on
this subject and most of the research was performed in
the past few years. As shown in Figure 1, trend lines
for both explanatory modeling and predictive modeling
are increasing, slowly but gradually. Trend lines would
increase even more if the year 2014 was removed from the
figures, as 2014 shows a big drop in the number of papers
published due to many papers not having been published
yet.
Second, trend lines in explanatory models seem to increase
faster than the trend lines in predictive models. Trend
lines for predictive models are often more flat, except for
the trend line in Figure 3, which can be ignored due to
a low amount of papers in this application. This implies
that over time, the percentage of explanatory models com-
pared to predictive models increases. However, as shown
in Figure 1, the increase of explanatory models is only
slightly. Therefore its difficult to predict in which direc-
tion the research will be going in the future.
Another point of interest is that predictive models are
slightly more recent than explanatory models. The first
paper about predictive models for predictive analytics in
the supply chain sector was released in 2004. In 2011 there
was a relatively big increase in papers regarding predictive
models. This could be explained by the fact that Shmueli
[66] released a paper about the difference between predic-
tive models and explanatory models in 2010. This can be
an indication that predictive models are still emerging.
Finally, research on both the applications monitoring and
inter-organizational collaboration started late, compared
to forecasting. This applies to both papers about predic-
tive models and papers about explanatory models. How-
ever, even though forecasting seems to be an older research
area there is still an upward trend going on in this area. It
is to be expected that the amount of research to all three
of the applications will continue to increase in the future.
6
4.6 Literature gaps
As mentioned earlier, literature regarding predictive ana-
lytics for supply chains is still in its infancy. There are
several literature gaps identified from the current litera-
ture. These literature gaps will be discussed on a high
level, for each of the categories discussed earlier: models,
methods, techniques and applications.
4.6.1 Models
There is a low amount of papers available on predictive
analytics actually giving predictions. Most of the papers
present a model, but do not apply new data to this model.
This is a major gap that does not only apply to predic-
tive analytics for the supply chain, but also for predictive
analytics in general [66].
4.6.2 Methods
Regarding the data collection methods, there are a low
amount of models based on data generated by simulation
as expected. It is unknown however if there are any conse-
quences regarding using simulated data or measured data.
4.6.3 Techniques
As shown in Table 3, well known techniques such as ma-
chine learning, support vector machines and system dy-
namics are clearly under-represented. This also applies
to other techniques such as naive Bayes, k-nearest neigh-
bours and decision trees, which are not even present at
all. It would be interesting to see if these techniques can
also be used in predictive analytics for the supply chain.
This could also cause new trends to appear in this research
area.
4.6.4 Applications
Besides the three major applications of predictive analyt-
ics for supply chains, predictive analytics can also be used
for other purposes. Under-researched applications as be-
come clear from Table 5 are price forecasting and predict-
ing product deficiency. Successfully applying predictive
analytics to one of these applications can be a cost-saving
process.
Besides the two mentioned applications above, there is
also a lack of research on customer sentiment data and
the usage of predictive analytics in human resources [74].
Data can be gathered in both of these areas and thus can
be potentially interesting for applying predictive analytics.
5. CONCLUSION
RQ1: What are the models, methods, techniques and ap-
plications used in predictive analytics for the supply chain?
There is only a limited amount of research available on
predictive analytics for supply chains. The relevant lit-
erature was mostly released in the past few years but is
growing rapidly. It is expected that the amount of litera-
ture released on a yearly basis will continue increasing the
upcoming years.
The literature was categorized based on models, methods,
techniques and applications. As shown in Table 1, the ma-
jority of the models are explanatory. Explanatory models
provide a model based on historical data but do not apply
new data to make predictions about the future, in contrary
to predictive models.
Data collection methods found were mostly based on mea-
sured data. These data collection methods include exper-
iments, case studies and surveys.
The most used techniques in predictive analytics for sup-
ply chains are time series, regression and neural networks.
However, the usage of the most used techniques in predic-
tive models seems to be lower than the usage of less used
techniques in predictive models, percentage-wise.
Applications of predictive analytics for supply chains are
mostly found in the categories forecasting (e.g. demand
forecasting, inventory forecasting, predicting bullwhip ef-
fect), inter-organizational collaboration (e.g. supplier se-
lection, predicting the influence of collaboration) and mon-
itoring (e.g. travel time prediction, order arrival time pre-
diction, predicting logistic costs).
Based on the categorization, some trends were identified.
As shown in the figures, the amount of papers on pre-
dictive analytics for the supply chain are still increasing.
The number of papers about explanatory models seem to
increase faster than the percentage of papers about predic-
tive models however, but papers about predictive models
are slightly more recent on average.
Research on the applications, monitoring and inter-organi-
zational collaboration seem to be newer compared to fore-
casting. The forecasting application is more mature com-
pared to the others, both based on time and on quantity.
RQ2: What are the research gaps of the literature found?
We have also identified research gaps. First, there is only
a low amount of papers that are actually making predic-
tions. Most of the papers discuss an explanatory model
instead of a predictive model. Second, there is a lack of re-
search in the applications other than the three mentioned
before. These applications include: price forecasting and
predicting product deficiency.
5.1 Limitations and future research
A major limitation is the lack of papers on this subject.
As mentioned earlier and shown in Figures 3 and 4, it is
hard to predict trends based on only a few papers on a
specific application. The reliability of some of the trends
is therefore questionable. It is expectable however that
research on this subject will increase as the trends show.
Future research could therefore provide a better overview
of the current state-of-the-art with an increase of relevant
papers.
6. ACKNOWLEDGEMENTS
A special thanks to supervisor Sjoerd van der Spoel for
providing feedback and assistance during the research.
References
[1] E. Aggelogiannaki and H. Sarimveis. Design of a novel
adaptive inventory control system based on the online
identification of lead time. International Journal of
Production Economics, 114(2):781792, 2008.
[2] F. T. Altekin and C. Akkan. Task-failure-driven re-
balancing of disassembly lines. International Journal
of Production Research, 50(18):49554976, 2012.
[3] A. Ancarani, C. Di Mauro, and D. DUrso. A hu-
man experiment on inventory decisions under supply
uncertainty. International Journal of Production Eco-
nomics, 142(1):6173, 2013.
[4] Association for Information Systems. Senior Scholars
Basket of Journals. http://aisnet.org/general/
custom.asp?page=SeniorScholarBasket, 2013. [On-
line; accessed on 12-03-2014].
[5] F. Barro et al. Stakeholder theory and dynamics
in supply chain collaboration. International Journal
7
of Operations & Production Management, 29(6):591
611, 2009.
[6] A. Bhattacharya, S. A. Kumar, M. Tiwari, and S. Tal-
luri. An intermodal freight transport system for op-
timal supply chain logistics. Transportation Research
Part C: Emerging Technologies, 38:7384, 2014.
[7] K. K. Boyer and M. T. Frohlich. Analysis of effects
of operational execution on repeat purchasing for het-
erogeneous customer segments. Production and Op-
erations Management, 15(2):229242, 2006.
[8] M. Cao, M. A. Vonderembse, Q. Zhang, and T. Ragu-
Nathan. Supply chain collaboration: conceptualisa-
tion and instrument development. International Jour-
nal of Production Research, 48(22):66136635, 2010.
[9] B. Chae, D. Olson, and C. Sheu. The impact of sup-
ply chain analytics on operational performance: a
resource-based view. International Journal of Pro-
duction Research, (ahead-of-print):116, 2013.
[10] S. Chan, W. Ip, and W. Zhang. Integrating fail-
ure analysis and risk analysis with quality assur-
ance in the design phase of medical product develop-
ment. International Journal of Production Research,
50(8):21902203, 2012.
[11] G. H. Chao. Production and availability policies
through the Markov Decision Process and myopic
methods for contractual and selective orders. Eu-
ropean Journal of Operational Research, 225(3):383
392, 2013.
[12] D. C. Chatfield. Underestimating the bullwhip effect:
a simulation study of the decomposability assump-
tion. International Journal of Production Research,
51(1):230244, 2013.
[13] K. C. Chatzidimitriou and A. L. Symeonidis. Data-
mining-enhanced agents in dynamic supply-chain-
management environments. Intelligent Systems,
IEEE, 24(3):5463, 2009.
[14] A. Chen and J. Blue. Performance analysis of demand
planning approaches for aggregating, forecasting and
disaggregating interrelated demands. International
Journal of Production Economics, 128(2):586602,
2010.
[15] L.-C. V. Cheng and E. E. Carrillo. Assessing supplier
performances under partnership in project-type pro-
curement. Industrial Management & Data Systems,
112(2):290312, 2012.
[16] G. Confessore, D. Corini, and G. Stecca. A computa-
tional method for pricing of delivery service in a lo-
gistics network. International Journal of Production
Research, 46(5):12311242, 2008.
[17] S. Dauz`ere-Per`es, A. Nordli, A. Olstad, K. Haugen,
U. Koester, M. Per Olav, G. Teistklub, and A. Reis-
tad. Omya Hustadmarmor optimizes its supply chain
for delivering calcium carbonate slurry to European
paper manufacturers. Interfaces, 37(1):3951, 2007.
[18] A. De Toni. Buyer-supplier operational practices,
sourcing policies and plant performances: results of
an empirical research. International Journal of Pro-
duction Research, 37(3):597619, 1999.
[19] J. Dejonckheere, S. Disney, M. R. Lambrecht, and
D. Towill. Transfer function analysis of forecasting in-
duced bullwhip in supply chains. International Jour-
nal of Production Economics, 78(2):133144, 2002.
[20] P. Doganis, E. Aggelogiannaki, and H. Sarimveis. A
combined model predictive control and time series
forecasting framework for production-inventory sys-
tems. International Journal of Production Research,
46(24):68416853, 2008.
[21] B. T. Ewing and M. A. Thompson. Industrial pro-
duction, volatility, and the supply chain. Interna-
tional Journal of Production Economics, 115(2):553
558, 2008.
[22] J. M. Field and L. C. Meile. Supplier relations and
supply chain performance in financial services pro-
cesses. International Journal of Operations & Pro-
duction Management, 28(2):185206, 2008.
[23] T. D. Fry and J. M. Donohue. Outlets for operations
management research: a DEA assessment of journal
quality and rankings. International Journal of Pro-
duction Research, 51(23-24):75017526, 2013.
[24] G. M. D. Ganga, L. C. R. Carpinetti, and P. R. Poli-
tano. A fuzzy logic approach to supply chain perfor-
mance management. International Journal of Pro-
duction Economics, 134(1):177187, 2011.
[25] C. A. Garcia Salcedo, A. Ibeas Hernandez, R. Vi-
lanova, and J. Herrera Cuartas. Inventory control of
supply chains: mitigating the bullwhip effect by cen-
tralized and decentralized internal model control ap-
proaches. European Journal of Operational Research,
224(2):261272, 2013.
[26] G. M. Gaukler, R. W. Seifert, and W. H. Hausman.
Item-Level RFID in the Retail Supply Chain. Produc-
tion and Operations Management, 16(1):6576, 2007.
[27] A. F. G uneri, T. Ertay, and A. Yucel. An approach
based on ANFIS input selection and modeling for sup-
plier selection problem. Expert Systems with Applica-
tions, 38(12):1490714917, 2011.
[28] M. V. Hermenegildo. Conferences vs. Journals in CS,
what to do? 2012.
[29] K. Hyndman, S. Kraiselburd, and N. Watson. Align-
ing capacity decisions in supply chains when demand
forecasts are private information: Theory and exper-
iment. Manufacturing & Service Operations Manage-
ment, 15(1):102117, 2013.
[30] V. Jain, S. Wadhwa, and S. Deshmukh. Supplier
selection using fuzzy association rules mining ap-
proach. International Journal of Production Re-
search, 45(6):13231353, 2007.
[31] S. Jaipuria and S. Mahapatra. An improved demand
forecasting method to reduce bullwhip effect in supply
chains. Expert Systems with Applications, 41(5):2395
2408, 2014.
[32] J. Jayaram, K.-C. Tan, and S. Nachiappan. Ex-
amining the interrelationships between supply chain
integration scope and supply chain management ef-
forts. International Journal of Production Research,
48(22):68376857, 2010.
8
[33] John Lamp. John Lamps IS Pages. http:
//lamp.infosys.deakin.edu.au/era/?page=
cfordet10&selfor=0806, 2010. [Online; accessed on
12-03-2014].
[34] P. Jula and R. C. Leachman. Long-and short-run
supply-chain optimization models for the allocation
and congestion management of containerized imports
from Asia to the United States. Transportation Re-
search Part E: Logistics and Transportation Review,
47(5):593608, 2011.
[35] R. Kapuscinski, R. Q. Zhang, P. Carbonneau,
R. Moore, and B. Reeves. Inventory decisions in Dells
supply chain. Interfaces, 34(3):191205, 2004.
[36] KDnuggets. Top research conferences in Data
Mining. http://www.kdnuggets.com/2013/11/top-
conferences-data-mining-data-science.html,
2014. [Online; accessed on 12-03-2014].
[37] W. Ketter, J. Collins, M. Gini, A. Gupta, and
P. Schrater. Detecting and forecasting economic
regimes in multi-agent automated exchanges. Deci-
sion Support Systems, 47(4):307318, 2009.
[38] W. Ketter, J. Collins, M. Gini, A. Gupta, and
P. Schrater. Real-time tactical and strategic sales
management for intelligent agents guided by eco-
nomic regimes. Information Systems Research,
23(4):12631283, 2012.
[39] T. Kim and C. H. Glock. On the use of RFID in
the management of reusable containers in closed-loop
supply chains under stochastic container return quan-
tities. Transportation Research Part E: Logistics and
Transportation Review, 64:1227, 2014.
[40] E. R. Kone and M. H. Karwan. Combining a new data
classification technique and regression analysis to pre-
dict the Cost-To-Serve new customers. Computers &
Industrial Engineering, 61(1):184197, 2011.
[41] J. Korpela and M. Tuominen. Inventory forecasting
with a multiple criteria decision tool. International
journal of production economics, 45(1):159168, 1996.
[42] A. W. Labib. A supplier selection model: a com-
parison of fuzzy logic and the analytic hierarchy pro-
cess. International Journal of Production Research,
49(21):62876299, 2011.
[43] H. C. Lau, G. T. Ho, and Y. Zhao. A demand forecast
model using a combination of surrogate data analy-
sis and optimal neural network approach. Decision
Support Systems, 54(3):14041416, 2013.
[44] R. C. Leachman and P. Jula. Congestion analysis
of waterborne, containerized imports from Asia to
the United States. Transportation Research Part E:
Logistics and Transportation Review, 47(6):9921004,
2011.
[45] C.-J. Lu and Y.-W. Wang. Combining independent
component analysis and growing hierarchical self-
organizing maps with support vector regression in
product demand forecasting. International Journal
of Production Economics, 128(2):603613, 2010.
[46] D. A. Menachof, B. J. Gibson, J. B. Hanna, and
A. E. Whiteing. An analysis of the value of supply
chain management periodicals. International Jour-
nal of Physical Distribution & Logistics Management,
39(2):145165, 2009.
[47] J. R. Meredith, M. D. Steward, and B. R. Lewis.
Knowledge dissemination in operations manage-
ment: Published perceptions versus academic reality.
Omega, 39(4):435446, 2011.
[48] Microsoft Research. Top conferences in data min-
ing. http://academic.research.microsoft.
com/RankList?entitytype=3&topDomainID=
2&subDomainID=7&last=0&start=1&end=100, 2013.
[Online; accessed on 12-03-2014].
[49] S. Moon, A. Simpson, and C. Hicks. The develop-
ment of a classification model for predicting the per-
formance of forecasting methods for naval spare parts
demand. International Journal of Production Eco-
nomics, 143(2):449454, 2013.
[50] D. C. Moosmayer, A. Y.-L. Chong, M. J. Liu, and
B. Schuppar. A neural network approach to pre-
dicting price negotiation outcomes in business-to-
business contexts. Expert Systems with Applications,
40(8):30283035, 2013.
[51] N. A. Mylonopoulos and V. Theoharakis. On site:
global perceptions of is journals. Communications of
the ACM, 44(9):2933, 2001.
[52] A. Oke and M. Szwejczewski. The relationship be-
tween UK manufacturers Z
aA inventory levels and
supply, internal and market factors. International
Journal of Production Economics, 93:151160, 2005.
[53] Y. Ouyang. The effect of information sharing on sup-
ply chain stability and the bullwhip effect. European
Journal of Operational Research, 182(3):11071121,
2007.
[54] Y. Ouyang and C. Daganzo. Characterization of
the bullwhip effect in linear, time-invariant supply
chains: some formulae and tests. Management sci-
ence, 52(10):15441556, 2006.
[55] Y. Ouyang and C. F. Daganzo. Discretization and
validation of the continuum approximation scheme
for terminal system design. Transportation Science,
40(1):8998, 2006.
[56] Y. Ouyang and X. Li. The bullwhip effect in sup-
ply chain networks. European Journal of Operational
Research, 201(3):799810, 2010.
[57] P. M. Panayides and Y. Venus Lun. The impact
of trust on innovativeness and supply chain perfor-
mance. International Journal of Production Eco-
nomics, 122(1):3546, 2009.
[58] M. Peng, Y. Peng, and H. Chen. Post-seismic sup-
ply chain risk management: A system dynamics dis-
ruption analysis approach for inventory and logistics
planning. Computers & Operations Research, 42:14
24, 2014.
[59] S. Rajagopalan. Impact of variety and distribution
system characteristics on inventory levels at US re-
tailers. Manufacturing & Service Operations Man-
agement, 15(2):191204, 2013.
[60] B. Rostami-Tabar, M. Z. Babai, A. Syntetos, and
Y. Ducq. Demand forecasting by temporal aggrega-
tion. Naval Research Logistics (NRL), 60(6):479498,
2013.
9
[61] B. Sahay and J. Ranjan. Real time business intelli-
gence in supply chain analytics. Information Man-
agement & Computer Security, 16(1):2848, 2008.
[62] B. Scholz-Reiter, J. Heger, and T. Hildebrandt. Gaus-
sian Processes for Dispatching Rule Selection in Pro-
duction Scheduling: Comparison of Learning Tech-
niques. In Data Mining Workshops (ICDMW), 2010
IEEE International Conference on, pages 631638.
IEEE, 2010.
[63] Scimago Lab. Journal Rankings on Decision Sci-
ences. http://www.scimagojr.com/journalrank.
php?area=1800, 2012. [Online; accessed on 12-03-
2014].
[64] A. Serenko and M. Dohan. Comparing the expert
survey and citation impact journal ranking methods:
Example from the field of artificial intelligence. Jour-
nal of Informetrics, 5(4):629648, 2011.
[65] B. Sezen. Relative effects of design, integration and
information sharing on supply chain performance.
Supply Chain Management: An International Jour-
nal, 13(3):233240, 2008.
[66] G. Shmueli et al. To explain or to predict? Statistical
Science, 25(3):289310, 2010.
[67] G. Shmueli and O. R. Koppius. Predictive analytics in
information systems research. MIS Quarterly, 35(3),
2011.
[68] A. Simroth and H. Zahle. Travel time prediction using
floating car data applied to logistics planning. Intelli-
gent Transportation Systems, IEEE Transactions on,
12(1):243253, 2011.
[69] J. S. Stonebraker, E. Gil, C. W. Kirkwood, and R. B.
Handfield. Impact factor as a metric to assess journals
where om research is published. Journal of Operations
Management, 30(1):2443, 2012.
[70] R. Tangsucheeva and V. Prabhu. Modeling and anal-
ysis of cash-flow bullwhip in supply chain. Interna-
tional Journal of Production Economics, 145(1):431
447, 2013.
[71] The University of Sydney. Journal Rankings
for Transport, Logistics and Supply Chain Man-
agement. http://sydney.edu.au/business/itls/
research/journal_rankings, 2013. [Online; ac-
cessed on 12-03-2014].
[72] Thomson Reuters. ISI Web of Knowledge -
Journal Citation Reports. http://admin-apps.
webofknowledge.com/JCR/JCR, 2014. [Online; ac-
cessed on 12-03-2014].
[73] M. Y. Vardi. Conferences vs. journals in computing
research. Commun. ACM, 52(5):5, 2009.
[74] M. A. Waller and S. E. Fawcett. Data science, pre-
dictive analytics, and big data: A revolution that
will transform supply chain design and management.
Journal of Business Logistics, 34(2):7784, 2013.
[75] S.-J. Wang, W.-L. Wang, C.-T. Huang, and S.-C.
Chen. Improving inventory effectiveness in RFID-
enabled global supply chain with Grey forecasting
model. The Journal of Strategic Information Systems,
20(3):307322, 2011.
[76] W. Wang, D. E. Rivera, and K. G. Kempf. Model APPENDIX
predictive control strategies for supply chain man-
agement in semiconductor manufacturing. Interna- A. LIST OF RESEARCH AREAS
tional Journal of Production Economics, 107(1):56
77, 2007. Information Systems

[77] T. Warren Liao and P.-C. Chang. Impacts of forecast, Computer Science: Artificial Intelligence
inventory policy, and lead time on supply chain inven- Computer Science: Interdisciplinary Applications
tory
aA Ta
numerical study. International Journal of
Production Economics, 128(2):527537, 2010. Engineering: Industrial
[78] J. Webster and R. T. Watson. Analyzing the past to Operations Research
prepare. MIS quarterly, 26(2):1323, 2002.
Supply Chain
[79] B. D. Williams, M. A. Waller, S. Ahire, and G. D.
Ferrier. Predicting retailer orders with POS and order Decision Sciences
data: The inventory balance effect. European Journal
Business
of Operational Research, 232(3):593600, 2014.
Management
[80] T. Wong, K. M. Law, H. K. Yau, and S.-C. Ngan. An-
alyzing supply chain operation models with the PC-
algorithm and the neural network. Expert Systems B. LIST OF SELECTED JOURNALS
with Applications, 38(6):75267534, 2011. The list of journals included can be found with the follow-
ing link:
[81] S. D. Wu, B. Aytac, R. T. Berger, and C. A. Arm-
bruster. Managing short life-cycle technology prod- https://docs.google.com/spreadsheet/ccc?key=0Ah_
ucts for Agere Systems. Interfaces, 36(3):234247, DiBPUohMUdFlGb05IWUdhRXpfdmxfMzlGaWJVT3c&usp=
2006. sharing
The table shows the journal titles as well as their appear-
[82] W. Xin-li and Z. Kun. Study on primary product ance in the journal rankings.
logistics: demand prediction based on neural network
theory. In Knowledge Discovery and Data Mining,
2010. WKDD10. Third International Conference on, C. LIST OF SELECTED CONFERENCES
pages 16. IEEE, 2010. The list of the conferences included can be found with the
following link:
[83] F. Zhang. An application of vector GARCH model in https://docs.google.com/spreadsheet/ccc?key=0Ah_
semiconductor demand planning. European journal DiBPUohMUdEJDNWZfaE5wZW9ScDFNN1lqamRHYkE&usp=
of operational research, 181(1):288297, 2007. sharing
[84] Z. Zou, T.-L. B. Tseng, H. Sohn, G. Song, and The table shows the conference titles as well as their ap-
R. Gutierrez. A rough set based approach to distrib- pearance in the conference rankings.
utor selection in supply chain management. Expert
Systems with Applications, 38(1):106115, 2011. D. RESEARCH QUERY
predict*
AND
(supply chain OR SCM OR logistics)
AND
(analy* OR big data OR data mining OR business
intelligence OR decision support)

10