Modeling knowledge reuse in technical support operations
Mauricio Uriona Maldonado’, Adriano Coser, Gregorio Varvakis
Knowledge Engineering and Management Post-Graduate Program.
Knowledge Engineering Department. Federal University of Santa Catarina (UFSC)
Campus UFSC, Bairro Trindade. Florianopolis, SC. CEP 88040-970. Brazil.
uriona@ieee.org ; coser@ egc.ufsc.br ; grego@ deps.ufsc.br
Abstract: There is an increasing concern on the part of corporate sector of the
importance to harness knowledge as their most valuable resource. The purpose of this
work is to identify the effects of knowledge reuse in service systems. In order to achieve
this, a system dynamics model of a Brazilian software-house’s technical support service
is presented, emphasizing on the use of knowledge bases and its effects over the service
system. It concludes that i) the model aids the designer in evaluating several aspects of
the system as well as its performance, including the effects of knowledge reuse and ii)
based on the simulation results, knowledge management enhances service system
performance.
Keywords: Service System Design, Knowledge Management, Knowledge base,
Knowledge reuse, System Dynamics
Introduction
Intangibility, simultaneity and non-stockability clearly differentiate service operations
from manufacturing ones (Gianesi & Correa, 1994). It has been suggested that service
operations depends intensively on the human capital involved. Usually, service systems
are based on a large number of interactions with both consumers and suppliers in which
value co-production is an inherent property (Tung & Yuan, 2007). According to Maglio
et al. (2006), “service systems are value-creation networks composed of people,
technology, and organizations”.
Cook et al. (2002) suggest that only with the understanding of the underlying principles
of human interactions, service design can be approached with the same depth and rigor
found in manufacturing operations. When the service system involves knowledge-
intensive activities, qualified human capital grows in importance, as well as the need to
strategically manage its large volumes of information and knowledge.
Service operations knowledge is crucial for bringing positive outcomes and superior
organizational performance. Knowledge Management is the discipline that addresses
those issues, by acquiring knowledge mainly from human sources, by codifying this
knowledge in order to be able to store it in knowledge-bases, and by re-using the
knowledge stored in the company’s processes (Uriona et al., forthcoming). .
' The authors wish to thank the National Council for Scientific and Technological Development (CNPq)
for supporting this research through the PEC/PG Program as a “Bolsista da CAPES/CNPq - IEL
Nacional - Brasil”
Knowledge bases are of extreme importance for service operations, since its human-
intensive nature, if correctly implemented and managed, therefore the need for them to
be considered under service design and production phases.
The design phase is essential for product development, establishing how the requisites
will be incorporated in the final manufactured goods or services. If the effects of
knowledge bases on a particular service system can be visualized and discussed on this
phase, several investments can be saved and the chance of a successful implementation
and results increases, by foreseen the positive/negative impacts knowledge reuse may
have over the system.
Thus, how can knowledge reuse and its effects be visualized in service operations
before service delivery/production? In manufacturing product design, it is common to
evaluate scenarios and alternatives related to product development. However, in service
design, due to its intangible nature, this evaluation is harder, making scenario-testing
more difficult. While, in goods production is common to build physical models that
materialize the conceived ideas, service design may use simulation techniques.
According to Banks (2000), simulation imitates the operation of a real-world process or
system during a period of time, based on the creation of an “artificial history” of the
system, where outcomes may help to infer real system operations. For Sheu et al. (2003)
simulation offers important advantages over mathematical tools, like value-ranges
flexibility in controlled parameters and real system behavior capture.
The purpose of this work is to identify the effects of knowledge reuse in service systems
by using System Dynamics as a support tool, which allows complex system simulation
through stock and flow metaphors.
In order to achieve this, a software-house’s technical support service is modeled using
System Dynamics. This activity has been described as knowledge-intensive and human-
based (Uriona, 2008).
The next section deals with the concepts of knowledge reuse. Third Section develops
service system design, highlighting the need to promote knowledge management
initiatives when it is considered a requisite. Fourth section develops the system
dynamics concepts used in the model. Model application is developed in Section 5,
considering a knowledge-intensive service system design in a software-house. Finally,
in Section 6, the conclusions of the paper are presented.
Knowledge Reuse as a feedback process
There is an increasing concern on the part of corporate sector of the importance to
harness knowledge as their most valuable resource. Many researchers have argued that
the capability to manage knowledge is the most important source of competitive
advantage (Nonaka and Takeuchi, 1995; Drucker, 1997).
The management of knowledge can be broken down in some phases, compounding the
so-called knowledge transformation cycle, Uriona et al. (forthcoming) proposes the
following phases: knowledge creation, formalization, store, share and use.
McElroy (2002) consolidates the cycle’s phases in the KMCI model (Exhibit 1). It
includes a series of feedback loops for organizational memory, beliefs, claim and
business-processing environment. McElroy also sustains that organizational knowledge
is held subjectively in the mind of individuals and groups and also objectively in
explicit forms.
Knowledge Production
Business Process Environment
Business Process Organizational Knowledge | Distributed
Behaviors of. |4———>> ‘Containers’ 41% Organizational
Feedback Interacting Agents |q—— + Artifacts & Codifications Knowledge
A (Knowledge Use) + Individuals and Teams
Internal/External Events
Exhibit 1: KMCI Model. McElroy (2002)
The knowledge production processes feeds the organization with new knowledge
through individual and group leaming, and with knowledge formulation, codification
and evaluation.
The knowledge integration processes introduces the new knowledge to its operating
environment, and via Single-loop Learning and Double-Loop Leaming (Argyris and
Schon, 1978) replaces old knowledge or re-starts the Knowledge production processes
for acquiring more knowledge.
As seen on Exhibit 1, Organizational Knowledge Containers, namely Knowledge Bases
have a direct impact on business process behavior - by using knowledge previously
acquired and codified - which in turn, feeds back on individual/group learning.
We infer that the more knowledge stored in artifacts compounding knowledge bases, the
more the impact on business performance, through process behavior.
Service System Design
Similarly to goods manufacturing, service operations are composed of several
components. However, these components are mainly non-physical, characterized by a
combination of processes, human competences and other resources (Goldstein et al.,
2002). In new service development or in service re-design, managers and designers
must make decisions with different levels of complexity about each component of the
service (Goldstein et al., 2002)
Service system design has been pointed out by Chase & Apte (2007), Hidaka (2006)
and Maglio et al. (2006) as a promising research field, considering also, the relevancy of
simulation and modeling techniques in helping analyzing these tasks.
Heineke & Davis (2007) discuss the relationship between the need for global service
expansions and the use of information and communication technologies with
geographically dispersed resources. These factors establish new challenges for service
managers and increase the importance of design and monitoring tasks for high quality
services.
Reinforcing the importance of investing in adequate HR management, constituting an
essential asset in service organizations, Dial (2007) points out that, in contrast to
manufacturing operations, services are highly dependent on operator's experience and
intuition, thus, having an inferior overall productivity than of manufacture activities.
The author suggests the adaptation and application of manufacture concepts and
methodologies in service operations in order to raise productivity indicators.
This paper - supported by the ideas exposed - recognizes the importance of knowledge
management in service operations, and points out the need to guarantee the necessary
resources in design phase. System Dynamics, as detailed in the next section, is explored
as a tool that seeks to help designers and to foreseen system’s behavior for each project
scenario, and more importantly, to analyze the effects of knowledge management
initiatives in the service operations.
System Dynamics
System Dynamics (SD) was developed by J. Forrester in 1961 (Forrester, 1989), as a
methodology for understanding complex systems behavior, through soft and hard
simulation. According to Sterman (2000):
“System Dynamics is a perspective and set of conceptual tools that enable
us to understand the structure and dynamics of complex systems. System
Dynamics is also a rigorous modeling method that enables us to build formal
computer simulations of complex systems and use them to design more
effective policies and organizations”.
It evolved from the application of control theory to the study of dynamic social systems.
Its premise is that the behavior of a complex dynamic system is the result of its structure
(causal relationships, feedback loops and time delays) (Sterman, 2000; Oliva &
Sterman, 2001).
Feedback loops are defined by information acquisition over system state and for actions
causing changes in that state. Its modeling involves accumulation processes (stocks) and
flows, as well as time delays and non-linear relationships. (Gonzalez & Dutt, 2007).
Sengir et al. (2004) discuss the importance of System Dynamics for behavior and
structure analysis in social systems. Feedback loops, differentiate system dynamics
from other approaches, by characterizing non-linear social relationships. Stocks and
flows of information, people and other resources allow the study of systems with high
levels of dynamic complexity and the study of timing issues in organizations.
Some of the advantages brought by this approach in modeling complex dynamic
systems are listed by Hollmann & Voss (2005): i) “stock and flow” diagrams provide
and intuitive vision above the structure of the system in study; ii) all the dependencies
and relationships are visualized graphically, facilitating the understanding of the
processes; iii) simulation tools, like iThink, allow model variables modification
interactively, in a so-called “control panel”, facilitating scenario-testing and analysis.
The Case Study
The Organization: An Overview
AltoQi Tecnologia em Informatica is located in Florianopolis, Brazil. The firm
develops, markets and supports software solutions for the construction market,
comprising architecture, structural analysis and engineering in concrete applications as
well as hydraulic, electric, fire sprinkler and gas layout projects.
Due to the complexity of its products and of its application domain, their costumers
generate large demands for specialized technical support. The firm maintains an
engineering and technical team that needs to be adequately trained in using the product
as well as in the project areas covered by the software products.
Nowadays the company counts with 11 workers in its technical support department,
mostly engineers, divided in three engineering areas: civil, electric and hydraulic. This
team answers the requests of approximately 17.000 users.
The services are made via telephone and email. On an average month, the team answers
around 81 calls per day with a mean of 13.8 minutes per call, likewise, around 32 email
replies with a mean of 27.6 minutes per email.
According to the Products and Services Department Manager, currently, the support
team has not been able to answer all of the customer demand. Daily, several calls are
lost and email requests seldom are answered on the same day due to the amount of
emails accumulated. These issues reduce customer satisfaction levels and also affect
team’s morale.
Another problem faced by this Department is tumover. When experienced workers
leave the team, they take with them knowledge that was acquired in their activities.
They are replaced with new workers that reduce overall performance, since these new
workers will usually be slower in their activities and will require more help from the
experienced colleagues until they leverage their knowledge.
The team’s manager and the company’s boardroom believed that an implementation of
a knowledge reuse strategy may aid in the problems explained. It is expected that
knowledge recovery and reuse applied to case-solving will serve to improve team
performance in future support services, by reducing service times and by improving its
quality. It is also expected that knowledge codified would diminish the effects of
tumover in the team, aiding in the training of new team members.
The System Dynamics Model
This section illustrates the use of a system dynamics model as a support tool for service
system design, emphasizing the possibility of simulating the knowledge management
effects over the system performance. The service analyzed from the practical field is a
technical support service, usually found in software developing companies. The data
supplied by AltoQi Tecnologia em Informatica allowed establishing the following
requisites:
= Opening requests are received via telephone and e-mail. Telephone support service has
priority over e-mail support service.
= The objective of the system is to reach zero non-attended telephone calls (zero
WaitingCalls) at the end of the day.
=> E-mail inbox is shared across the attendants and it is also desirable to be zero (0)
WaitingEmail at the end of the day.
= Knowledge management: at the end of a service, the attendant responsible should feed
the knowledge-base, aiming at making forthcoming services more agile.
It is beyond the scope of this paper to define the knowledge management strategies and
tools used in the company. The term “knowledge base” is used in this paper to represent
a knowledge repository that grows as the feeding process goes on. It is expected that
this knowledge repository will facilitate future attendances, if supported by adequate
knowledge representation, retrieving and reusing techniques.
Besides improving service system performance, the establishment of a knowledge base
must facilitate rookies training, reducing the expected effects of tumover. In
knowledge-intensive activities, such as technical support services, this aspect is of
fundamental importance, since attendants must accumulate a large volume of
knowledge - regarding functional characteristics of the software as well as technical
knowledge from the application domain - in order to successfully execute the activity
The example used aims to demonstrate that the structure needed to promote knowledge
management initiatives is a part of service system design and that - as the rest of its
components - it represents operational and financial costs that must be compensated or,
preferably, being overcome by the benefits brought by its use. The developed System
Dynamics model helps the service designer in evaluating these cost-benefit ratios by
considering several demands versus capacity scenarios.
The Model
Exhibit 2 shows the macro-model of the service system in study. The modules that
constitute it are: Phone Support, E-mail Support, Workforce, Knowledge Base and
Performance Measurement.
Exhibit 2: Macro-Model of the technical support service
Exhibit 3 shows the system dynamics model in iThink language. The five areas of the
model are described below.
=> Phone Support: the number of calls received daily is regulated by the IncomingCalls
inflow. The stock CallsToAnswer is emptied by AnsweringCalls and LosingCalls
outflows, this last-one represents excessive demand. The phone calls effectively
answered are accumulated in the CallsAnswered stock, that serves as a feeding source
for the KnowledgeBase stock.
= E-mail Support: IncomingEmail flow feds daily the EMailInbox stock. Differently
from phone-calls, e-mail inbox doesn’t necessarily empties-out, since it doesn’t count
with an outflow other than ReplyingEmail. The amount of emails answered each day
depends on the remaining time the attendants have after answering all of the phone-
calls. Same as the CallsAnswered stock, the EmailReplied stock feeds the
KnowledgeBase stock.
= Workforce: The stock AttendantsInService varies depending on the quantity of calls
and emails to answer. This policy helps to maximize the use of the workers as well as
their time, both reflected on the TotalCosts stock. The HiringRate parameter depends
exclusively on the amount of calls, reflected on the AnsweringCalls/CallsToAnswer
ratio. On the other hand, the TimeAvailable stock is shared in order to answer calls and
reply to e-mails. This depends initially, on the quantity of CallsToAnswer, if there is
less than 5 calls to answer, immediately an amount of time becomes available to answer
emails, however, if e-mail inbox has a quantity larger than 50 emails, an extra time
becomes available to answer emails, in order to reduce the amount of emails in
EmailInbox, leaving less time to answer phone calls.
= Knowledge Base: The KnowledgeBase stock is feed-up via knowledge processing,
which uses CallsAnswered and EmailReplied. The KnowledgeProcessingCall and the
KnowledgeProcessingEmail need a minimum amount of data and information stored on
CallsAnswered and EmailReplied in order to be useful, a 1000 for each one. The
KnowledgeBase then, is used to reduce the TimePerEmail and the TimePerCall
parameters.
= Performance Measurement: This area helps to measure key-performance indicators
for the support service. These are: WaitingCalls that represent the amount of calls left in
the waiting line each day; WaitingEmail that represents the amount of emails left in the
EmailInbox without replying each day and TotalCosts that represent the costs due to
workforce demand.
oo Powe spor a oo Wortoce Fy
CCaleTonnener — AnswerneCaks——Cateanswered
Losngcals
Cy
og Eralsipen
Emalfent
S emai
bs rotacosts D
og Krewe Base @ [COD _retomanceneasuenert 6
O
epg mai eal, Ccaelatagne
TeonigafeProcersingemal
eMaiRepted
Exhibit 3: Technical support service model
Dynamic complexity of the model is also regulated through a series of variables
tepresenting several activities. AverageCalls and AverageEmails represent the amount
of calls and emails that arrive each day. Both are statistically represented by Poisson
Probability functions.
Scenario 1: Pilot Test
Scenario 1 served as a pilot test for the model, in order to validate if the output data
from the model was effectively representing the firm’s reality. This scenario does not
consider a periodical use of a Knowledge Base. Exhibit 4 shows the simulation
outcomes for a 365 day period
&@ 1: calstoanswer
rf 300%
183.00 274.00 365.00,
Page 2 Days
» ?
B®» evatinvox
300 | |
1.00 92°00 183.00 274.00 365.00
Page 1 Days
Exhibit 4: System behavior considering low stock of knowledge
As shown in Exhibit 4, the amount of calls to answer has a fluctuant behavior, with
peaks near 150 calls per day. The situation with email inbox is similar with peaks over
150 emails per day. When CallsToAnswer reaches zero, attendants boost their email
replying, visualized in Exhibit 3 between the 190" and 230" days of the simulation
period, when the amount of emails to reply reduces drastically due to the near-zero level
of CallsToAnswer.
This behavior can also be seen at different moments of the simulation in a lesser scale,
sustaining the fact that when CallsToAnswer reach levels close to zero, the attendants
have more time available to reply emails.
Exhibit 4 also shows that the average amount of non-attended calls - WaitingCalls - per
day is 69,09, this means that on an average day, 69,09 calls will be in the waiting line.
For emails the WaitingE mail average is 110,38.
In order to test and validate the model, the outputs of CallsAnsweredPerDay and
EmailRepliedPerD ay were confronted with actual data of the firm.
Variables
Calls answered per da
Emails replied per day
According to Table 1, the errors between actual data and simulation outputs are less
Outputs of Actual Data Divergence
Simulation (%)
y 78,64 81,31 3,28
32,02 32,81 2,40
Table 1. Validity Test.
than 4%, thus, making the model reliable for further experiments.
Scenario 2: Knowledge Base considered
Once the model was tested and that its reliability was confirmed, the Scenario 2
considers the use of a Knowledge Base as a service time reducing element.
Exhibit 5 shows the results of this scenario. In this Scenario, the Knowledge Base starts
being used after the first 1000 calls answered and 1000 emails replied (approximately
the first 90 days of the simulation).
& 1: caistosnswer
1 300
92,00
365.00
1 t
100
Page 2
® »: enaiindox
f 300.
1 150
1 0
00
paged
IN 2
92.00
274,00
365.00
Exhibit 5: Syste
m behavior considering high stock of knowledge
10
As shown in Exhibit 5, the amount of CallsToAnswer is lesser than in Scenario 1. This
is due to the use of a Knowledge Base, which allows to reuse the knowledge
incorporated in the technical support. There is also a 120 days period approximately
when the amount of emails to reply reduces significantly below the 150 emails limit, at
the same time when CallsToAnswer reduces its backlog.
Exhibit 5 also shows that the average amount of non-attended calls - WaitingCalls - per
day is 63,31, this means that on an average day, 63,31 calls will be in the waiting line.
For emails the WaitingEmail average is 104,59.
Discussion
Table 2 illustrates the summary of the results in both simulations.
Key-Performance Indicators Scenario1 Scenario 2 %
Waiting Calls (Number of calls) 69,09 63,31 -8,36
Waiting Email (Number of emails) 110,38 104,59 -5,24
Calls answered per day 78,64 82.38 44,75
Emails replied per day 32,02 31.14 -2,74
Table 2. Summary of the results,
The results have indicated that KnowledgeBase do have significant influence on the
service quality in knowledge-intensive service systems.
As shown on Table 3, Scenario 2 presents reductions for both waiting calls and waiting
emails, for the first one around 8% and for the latter around 5%.
This reduction can be explained by using McElroy’s KMCI Model, where the use of
artifacts and codification tools able changes in service and process behavior, working as
a feedback loop that reinforces the organizational learning process.
In other words, workers increase their service speed and are capable of answering more
calls and replying more emails, by using a knowledge base as a stock of knowledge
about past services.
Table 2 also presents the results for calls answered and emails replied. For the first one,
there is an increase of calls answered of 5% approx., this increase means that technical
operators are able to answer 4% more calls per day, once the Knowledge Base is
implemented.
For emails replied, there is a reduction of 3% approx., although, this reduction implies
on lesser emails answered, the percentage of difference can be understood as a indirect
effect of the models random variability. This difference would represent numerically
around one (1) email less answered per day.
11
Conclusions and Future Research
Flexibility in service operations and design-ability are the fundamental requirements of
modem day service sectors. In today’s customer driven market, these requirements are
of paramount importance more than ever before.
This paper has tried to demonstrate the usefulness of simulation techniques, such as
System Dynamics, in service system design. Specifically, it aimed to analyze the effects
caused by knowledge reuse in a technical support service.
It concludes that knowledge management implementations should be analyzed earlier in
the design phase, supported by simulation techniques for scenario-testing and
evaluation. The System Dynamics model was developed using real data of a software-
house. This data was used to create two different scenarios, showing the importance of
knowledge management initiatives in knowledge-intensive activities.
Even though this paper culminates in the recommendation of using simulation
techniques for service system design in the studied field, it calls for future extension of
this research into the specific details of knowledge conversion, i.e. the SECI model of
Nonaka and Takeuchi (1995) so as to facilitate the storage of knowledge in the
KnowledgeBase; as well as processes such as organizational forgetting and un-leaming
that outflows the KnowledgeBase stock. There is also enough scope to add new
variables into the model (e.g. organizational culture factors, motivation issues,
demographic influences, etc.), which influence service system design.
References
Argyris, C. and Schon, D. A. (1978). Organizational Learning: A Theory of Action Perspective,
Reading, MA: Addison-Wesley.
Banks, J. (2000), “Introduction to simulation”, Proceedings of Simulation Conference, Orlando, FL,
USA, pp. 9 - 16.
Chase, R, B., Apte, U. M. (2007), “A history of research in service operations: What's the big idea?”,
Joumal of Operations Management, Vol. 25, Issue 2, pp. 375 - 386.
Cook, L. S., Bowenb, D. E., Chase, R. B., Dasuc, S., Stewart , D. M., Tansik , D. A. (2002), “Human
issues in service design”, Journal of Operations Management, Vol. 20, Issue 2, pp. 159 - 174.
Drucker, P. (1997). Sociedade Post-Capitalista, Sao Paulo: Pioneira.
Diao, Y. (2007), “Using control theory to improve productivity of service systems”, IEEE
Intemational Conference on Services Computing, pp. 435 - 442.
Forrester, J. (1989), “The Beginning of System Dynamics’, International Meeting of the System
Dynamics Society, Stuttgart, Germany.
Gianesi, I. G. N., Corréa, H. L. (1994), “Administracao estratégica de servicos: operacées para a
satisfacao do cliente”, Editora Atlas S.A., Brasil, ISBN 85-224-1152-2.
Goldstein, S. M., Johnston, R., Duffy, J., Rao, J. (2002), “The service concept: the missing link in
service design research?” , Journal of Operations Management, Vol. 20, Issue 2, pp. 121 - 134.
Gonzalez, C., Dutt V. (2007), “Learning to Control a Dynamic Task: A System Dynamics C ognitive
Model of the Slope Effect”, Proceedings of the Eighth Intemational Conference on Cognitive
Modeling, pp. 61 - 66.
Heineke, J., Davis, M. M. (2007), “The emergence of service operations management as an academic
discipline”, Jounal of Operations Management, Vol. 25, Issue 2, pp. 364 - 374.
12
oss, J. (2005), “Modeling of decentralized energy supply structures with "system
International Conference on Future Power Systems.
Karwan , K. R., Markland, R. E. (2006), “Integrating service design principles and information
technology to improve delivery and productivity in public sector operations: The case of the
South Carolina DMV”, Journal of Operations Management, Vol. 24, Isuue 4, pp. 347 - 362.
Maglio, P.P.; Srinivasan, S.; Kreulen, J.T.; Spohrer, J. (2006). Service Systems, service scientists,
SSME and innovation. Communications of the ACM, v.49, n.7, p.81-85.
McElroy, M.W. (2002). Social Innovation Capital. Journal of Intellectual Capital. Vol. 3, Issue 1, pp.
30-39.
Nonaka, I.; Takeuchi, H. (1995). The knowledge-creating company: how Japanese companies create
the dynamics of innovation. New Y ork: Oxford University Press.
Oliva, R., Sterman, J.D. (2001). Cutting corners and working overtime: Quality erosion in the service
industry. Management Science. Vol. 47, n. 7, p. 894-914.
Sengir, G. H., Trotter, R. T., Briody, E. K., Kulkami, D. M., Catlin, L. B., Meerwarth, T. L. (2004),
“Modeling relationship dynamics in GM's research-institution partnerships’, Joumal of
Manufacturing Technology Management, Vol. 15, Issue 7, pp. 541 - 559.
Sheu, C., McHaney, R., Babbar, S. (2003), “Service process design flexibility and customer waiting
time”, International Journal of Operations & Production Management, Vol. 23, Issue 8, pp. 901 -917.
Sterman, J.D. (2000). Business Dynamics. Systems Thinking and Modeling for a complex world.
Boston. Mc Graw Hill Higher Education.
Tung W., Yuan S. (2007), “Design : An Intelligent Design Framework for Service
Innovation", Proceedings of the 40th Hawaii Intemational Conference on System Sciences, pp. 64.
Uriona Maldonado, M.; Soratto, A.N.; Giugliani, E.; Dos Santos, N. and Steil, A.V. (forthcoming).
Towards an Organizational Learning and Knowledge Management integrative view: in search of a
common ground. Intagible Capital.
Uriona Maldonado, M. (2008). Impact Analysis of knowledge-creation and knowledge-transfer
policies in knowledge-intensive business processes. A System Dynamics Model. M.Sc.Thesis.
Florianopolis, Brazil: Federal University of Santa Catarina.
13
