Description
Collaborative Decisions within Business Intelligence Context A GDSS prototype
Abstract—In this work we present a prototype for a web based
Group Decision Support System, which can be integrated within a
firm’s Business Intelligence (BI) architecture to support decisions in
small collaborating teams. It is based on web technology and can be
used in asynchronous mode from group members. It implements a
multicriteria methodology for classification decisions where
aggregation of members’ preferences is performed at the parameter
level. Initially, a set of parameters defined by stakeholders is
proposed to the group by group facilitator. Next, each group member
evaluates the proposed parameter set and expresses her preferences in
numeric or linguistic format. Individual preferences are aggregated
by appropriate operators, and a group parameter set is produced
which is used as input for the classification algorithm. NeXClass
multicriteria classification algorithm is used for the classification of
alternatives, initially at a training set of alternatives and later at the
entire set. Finally, group members evaluate results, and consensus as
well as satisfaction metrics are calculated. In case of low acceptance
level, problem parameters are redefined by group members and
aggregation phase is repeated. The system has been utilized to solve
real world group classification problems, especially within the BI
environment of a commercial bank, supporting mainly financial
decisions. Empirical findings from the GDSS application have been
evaluated and enhancements have already been incorporated in order
to improve existing functionality and provide additional features.
Regarding the overall approach, findings provide evidence that it is
appropriate for similar decision problems in numerous business
environments, and the GDSS can be a valuable tool for enhancing a
BI framework with advanced decision capabilities.
Keywords—Business intelligence, Group decision support,
multicriteria decisions.
I. INTRODUCTION
OLLOWING advances in Information Technology the
majority of processes and decisions in large firms and
organizations can be characterized today as data driven. The
ability of acquisition and organization as well as development
and diffusion of knowledge has become a critical factor for
market performance and firm viability. Moreover, the amount
of information gathered from traditional and novel sources
such as customers, Internet and information systems, is
excessively increasing requiring additional organizational
effort. As a consequence, increasing complexity at the
G.A.R. is with the Higher Technological Educational Institute of Athens,
Athens, Greece (e-mail: [email protected]).
N.V.K. is with the Hellenic Military Academy, Faculty of Military
Sciences, Division of Mathematics and Engineering Sciences, Vari, Greece
(corresponding author’s contact details: +306932121552 e-mail:
[email protected]).
knowledge level has led to additional needs for advanced
decision support. Provision thus of appropriate decision
support tools at managerial as well as at operational level is
critical for efficient performance, and moreover it offers firms
a strong advantage against the rest of the competitors within
the domain [1], [2], [3].
One of the Information Technology directions that aims to
support firms handle the above complexity is Business
Intelligence (BI). BI provides a set of methods, processes and
tools to support firms’ decisions through intelligent
exploration, integration, aggregation and analysis of data from
various resources. In brief, the origins of BI tools can be
traced at the early developments of Decision Support Systems
(DSS), while current needs for advanced intelligent decisions
due to information complexity has led to a massive
development of BI systems with DSSs’ being a subset of BI
domain. Within a BI architecture, a DSS stands on top of BI
tools, utilizing aggregated information provided though them,
to assist decision makers optimize their decisions.
Additionally, since most of the decisions within firms today
require the participation of a group of decision makers, it is
critical to provide tools to assist them within the context of a
BI framework [4], [5], [6].
Group decision support has received significant attention
from researchers due to its potential application to various
business domains. Research in decision support systems
targets towards supplying decision makers with appropriate
tools to assist them in optimizing their decisions. Since a
decision support system has to reflect decision makers’
preferences or decision model, building an appropriate Group
Decision Support System (GDSS) is not a straightforward
process. Moreover, a number of issues have to be considered
such as individuals’ preference modeling, negotiation
protocols and coordination, to mention a few.
Several methodologies and tools have been developed in
order to support groups, ranging from collaborative techniques
to negotiation ones, depending on whether group members
share a common goal or support individual goals.
Technologies acquired for developing a GDSS tend to follow
advances in Information Technology, resulting in recent
advanced systems. Incorporation of web technologies
nowadays, for example, can support collaboration features
which could not be implemented in the very early GDSSs.
However, from literature analysis there are very few
approaches that combine BI concepts and Group Decision
Collaborative Decisions within Business
Intelligence Context: A GDSS prototype
George A. Rigopoulos, Nikolaos V. Karadimas
F
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 70
Support Systems (GDSS), which is the main contribution of
our work [7], [8], [9].
Following the above, we developed a structured
methodology which is based on multicriteria analysis and
supports group classification decisions and a GDSS which
implements it. In this paper we present the methodology and
architecture along with a prototype.
The overall architecture and development of the GDSS was
based on web technology in order to be easily integrated
within an existing BI infrastructure. We followed a layered
approach, implementing concepts from Service Oriented
Architecture (SOA), aiming to provide a subset of the
functionality of the GDSS in terms of Web Services. The
proposed GDSS can be part of an existing BI architecture
within a firm or an organization, gathering input from several
BI subsystems to integrate them into a decision support
framework.
The GDSS has been deployed in real business environment
supporting mainly financial decisions. Regarding the overall
methodology, findings provide evidence that it is a valid
approach for similar decision problems in numerous business
environments within a BI architecture.
The structure of the paper is as follows. Following the
introduction, relevant background information in BI and
decision support as well as group decisions is presented. Next,
we present the integrated group decision multicriteria
methodology implemented by the GDSS. Section 4 presents
the architecture of the developed GDSS mentioning key
features of the system. Finally, the conclusion summarizes the
work.
II. BACKGROUND AND RELEVANT WORK
A. BI and decision support
BI is generally defined as “a term to describe leveraging the
organization’s information assets for making better business
decisions” [10]. Intelligence in BI is often defined as the
discovery and explanation of hidden, inherent and decision-
relevant contexts in large amounts of business and economic
data. BI consists today one of the fastest developing domains
in Information Technology. It is widely assumed that in the
near future BI systems integration with CRM (Customer
Relationships Management) and ERP systems (Enterprise
Resource Planning) will provide firms a strong competitive
advantage, enhancing quality of managerial decisions [11], [5].
In general, BI systems combine data from internal
information systems of a firm and integrate with data coming
from the environment such as. statistics and financial
databases, to provide adequate and reliable up-to-date
information on different aspects of firms’ activities [5]. The
use of BI tools is popular in industry [12], [13], indicating the
firms’ growing needs to handle the vast aggregation of
information orienting from business documents and data,
including business forms, databases, spreadsheets, e-mails,
articles, technical journals and reports, contracts, and web
documents. Distinction between knowledge management and
BI is not always clear [14], although knowledge and content
management technologies search, organize and extract value
from information sources, while BI focuses on the same
purpose, but from a different scope.
BI is mainly targeting in advancing decision making
utilizing data warehousing and online analytical processing
techniques (OLAP), collecting relevant data into repositories,
where organized and validated can be further available for
decision making. In general, business data are extracted,
transformed and loaded from various transactional systems
into a data warehouse after data cleansing processes and
multidimensional models are created to support drill down and
roll up analyses. A number of vendors provide tools and
platforms for such operations and advanced end user
functionality to support large amount of data [15].
From the above, the linkage between BI and decision
support within firms is evident. Moreover, BI origins can be
traced back at the early data-driven DSS approaches [6]. Later,
BI term was promoted and used as an umbrella term “to
describe a set of concepts and methods to improve business
decision making by using fact-based support systems”.
Although BI is sometimes used instead of the term of
executive information systems (EIS), in general BI systems
can be defined as data-driven DSSs. With the advent of
Internet, BI vendors shifted their BI solutions towards web
technologies and enterprise BI portals emerged [16].
B. Group decisions
Group decision making has become an essential component
of both strategic planning and everyday operations for the
majority of today’s organizations and enterprises. Since
complexity of business environment requires sufficient
knowledge from a wide range of domains, contribution of a
team of experts with key skills is the only way to achieve
efficiency in decisions. In order to support groups’ needs,
various researchers work on developing tools and
methodologies, ranging from collaboration technologies to
decision support systems [7], [8].
However, group decisions are quite more complex
compared to single decision making, since a number of
contradicting factors are involved such as individuals’ personal
opinions, goals and stakes resulting in a social procedure,
where negotiation and strategy plays a critical role.
Group decision making in real business environments raises
also some issues such as:
• Conflicting individual goals,
• Not efficient knowledge,
• Validity of information,
• Individuals’ motivation.
Despite the inherent complexity, within a group decision
making setting, a member is able to express personal opinions
and suggest solutions from a personal perspective. In addition,
negotiation and voting advance efficiency of decisions and
increment acceptability and adoption since all participants
have contributed to the result, smoothening thus any disputes.
In general, group members can be motivated by individual
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 71
perceptions to work within the group either towards
collaboration or towards competition. While in the first case,
members express similar opinions and goals, in the second one
they state opposite opinions. Although collaborative teams
work towards a common goal contradictions may also occur.
Some key techniques that have been acquired in order to
facilitate group work and decision include:
• Brainstorming,
• Nominal group technique,
• Delphi method,
• Voting,
• Multicriteria analysis.
C. Group decisions and multicriteria analysis
Group decision support is a subset of the more extended
research area of group support and negotiation. Group
decision support is an active research topic and existing
literature is quite extensive covering business as well as social
issues. Limiting the scope of relevant literature to integration
of multicriteria analysis within group decisions, we performed
a detailed survey focusing on relevant approaches, and
especially on developed Group decision support systems. An
extensive review of multicriteria analysis integration within
GDSSs is presented by Rigopoulos [7], [8], [9].
In general, multicriteria analysis can be incorporated as a
method to model preferences and facilitate decision making
within a group of decision makers. Modeling under a
multicriteria setting can be formulated under two major
directions:
1) In the first approach, individual multicriteria models are
developed, which capture individuals’ preferences. Each
group member formulates a multicriteria problem defining
the parameters according to her preferences and solves the
problem getting an individual solution set. Next, the
separate solutions are aggregated by aggregation operators
providing thus the group solution.
2) In the second approach, one multicriteria model is
developed for the entire team. Each group member
provides a set of parameters which are aggregated by
appropriate operators, providing finally a group parameter
set. Upon this set the muticriteria method is applied and
the solution expresses the group preference.
Each approach poses both positive and negative aspects
depending on the aggregation operation, which is followed.
III. PROPOSED GROUP DECISION METHODOLOGY
The main objective of our work is to provide support to a
group of decision makers in classification problems. The
problem refers to the assignment of a set of alternatives in a
number of predefined non-ordered categories, according to
their performance on a set of evaluation criteria. For this
reason, we have developed a structured group decision
methodology, which is based on the following principles:
1) The decision group is a small homogeneous team of
collaborating decision makers. Although the methodology
can be extended to large decision teams, our approach is
based on collaborative teams, which target towards
maximizing consensus. Non-collaborative teams require a
negotiation-based approach, which is out of scope of the
present methodology.
2) A facilitator coordinates the entire decision process. The
entire group decision process is coordinated by a
Facilitator. Usually, in group decision making a
negotiation phase takes place at the preliminary steps of
the decision problem formation. During this negotiation,
which can be structured or not, basic parameters are
defined. Since our methodology is not focusing on group
formation procedure and initial negotiations, we consider
that a preliminary negotiation step has already been
executed, possibly by utilizing brainstorming technique,
between stakeholders, and the outcome of this process is
an initial proposal of parameters. This set is expressed
from Facilitator as the initial proposal upon which group
members will express their preferences. Facilitator guides
the entire process in order to produce efficient and timely
results.
3) Decision problem is structured or semi structured. The
team solves a structured classification problem
contributing their preferences. Non structured problems
are out of scope.
4) Multicriteria analysis is utilized for the classification. For
the classification problem we utilize multicriteria analysis
which provides appropriate support to similar problems.
Following the above principles, we developed a group
decision methodology which is separated in the following
major phases:
A. Problem initiation
. In this phase Facilitator initiates the decision problem,
defining all appropriate parameters. The parameters are related
to the specific multicriteria methodology, and refer to criteria,
alternatives and categories. In details:
1) Basic parameters. Initially, Facilitator defines a number of
basic parameters, related to classification problem such as
the number of group members, the number of categories,
the number of criteria, and to results assessment such as
the consensus, satisfaction and acceptance levels. These
levels define minimum required levels for the group
decision. In case they are not satisfied, a second round is
executed with modification of individual preferences.
2) Members. Facilitator defines the group members assigning
all necessary contact details.
3) Categories. Facilitator defines the set of categories for the
classification of alternatives.
4) Evaluation criteria. Facilitator defines the set of
evaluation criteria according to problem requirements.
5) Criteria weights. Facilitator defines the criteria weights.
6) Alternatives. Facilitator defines the set of alternatives for
classification, and defines their performance on the
evaluation criteria.
7) Entrance thresholds. Facilitator defines appropriate
entrance thresholds for each category.
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 72
8) For each threshold Facilitator defines preference,
indifference and veto thresholds similar to ELECTRE TRI
method.
9) Training set. Facilitator defines a subset of alternatives as
training set, in order to evaluate parameters’ accuracy.
After the initiation of the parameters, Facilitator
communicates through the GDSS with group members
informing them about the problem and asking them to submit
their preferences.
B. Aggregation of individual parameters
In this phase group members express their preferences on
the proposed parameter set. Member preferences are expressed
in numeric values and linguistic preferences. For the
aggregation of numeric values we utilize the Social Judgment
Scheme (SJS), while linguistic terms are aggregated in terms
of an Ordered Weighted Averaging Operator (OWA) [18],
[19].
Aggregation of member preferences is executed for the
following parameters
1) Criteria. Group members express their acceptance on each
proposed criterion in a five point linguistic scale and their
preferred weight in numeric value.
2) Alternatives. Group members express their acceptance on
alternatives’ performance or submit their preference in
numeric value.
3) Categories. Group members express their acceptance on
each category definition, and submit their preferences on
category thresholds in numeric value. .
Facilitator proceeds with validation of members’ input and
aggregates the values. Parameters with low acceptance level
are marked and are subject to review if final results are not
acceptable by group members.
C. Application of multicriteria classification algorithm
After the aggregation of individual members’ parameters a
group parameter set is created and NeXClass algorithm for
multicriteria classification is applied on this group parameter
set [17]. NeXClass algorithm classifies an alternative to a
specific category with respect to alternative’s performance to
the evaluation criteria, considering a set of alternatives, a set of
predefined non-ordered categories and a set of evaluation
criteria. In more details the algorithm works as follows:
1) For each category decision maker defines an entrance
threshold using available information. This threshold
represents the minimum requirements for an alternative in
terms of performance on the evaluation criteria in order to
be included in this category.
2) Decision maker defines alternatives’ performance on the
evaluation criteria.
3) For each alternative an excluding degree is calculated for
every category threshold, based on outranking relations,
following a similar approach to ELECTRE TRI method.
4) Next, the fuzzy excluding degree of an alternative over a
category is calculated.
5) Assignment to a category is based on the rule which states
that alternative is assigned to the category for which the
excluding degree over the entrance threshold is minimum.
Application of NeXClass classification algorithm is
executed through the following steps
1) Training set classification. Classification algorithm is
initially applied to the training set initially, as it has been
defined by group members. Classification is executed by
Facilitator, and group members are informed to assess the
results.
2) Evaluation of results. Each member expresses her
preference on the results in a five point linguistic scale,
and in case of low acceptance level, Facilitaror executes a
second round of parameter definition from members in
order to calibrate the model. When training set
classification is acceptable, Facilitator proceeds with the
classification of the entire set of alternatives. In case of
low acceptance level after the second round, Facilitator
terminates the process in order to revise the problem with
stakeholders.
3) Training set classification. Classification algorithm is
finally applied to the entire set by Facilitator, and group
members are informed to assess the results.
D. Results assessment
Group members assess the results expressing their
preference in a five point linguistic scale. In case of low
acceptance level, Facilitator reruns the model, requesting
modifications from members.
IV. GROUP DECISION SUPPORT SYSTEM
A. Overview
A prototype GDSS was developed to implement the
aforementioned methodology. The design of the GDSS was
based on the following requirements:
1) Collaboration. The GDSS should promote collaboration
between group members by appropriate functions. Group
members for similar decision problems, can be selected ad
hoc without any prior collaboration. The GDSS should
thus promote the feeling of a common goal to members
minimizing thus individual goals.
2) Communication. Since business operations may span over
several locations, members can be located separately.
Communication thus between facilitator and group
members should be efficient enough in order to provide
results in a timely way. The GDSS should thus provide
appropriate communication tools.
3) Anonymity. Although anonymity poses some negative
issues, it encourages members express their preferences
without restrictions or external influence. For this reason,
the GDSS should support anonymity at presentation level.
4) Asynchronous operation. Different time zones and
locations of today’s business operations require
asynchronous operation and decision making. The GDSS
should thus provide asynchronous operation efficiently.
Considering the above requirements, a layered approach
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 73
which can be easily deployed in an existing BI infrastructure
was selected for the GDSS architecture (Fig. 1). BI
architecture components such as transactional data source
systems and data warehouses can be easily connected and
integrated with the GDSS, being the sources of input data. In
the same way, GDSS output can be deployed in firm’s BI
systems advancing firm’s knowledge. Regarding the
technology utilized, GDSS modules have been entirely
developed in Java language using JCharts library for chart
preparation. Apache web server is used to host the entire site
with Tomcat as servlet container and Tomcat Axis used for the
deployment of web services. Data layer has been implemented
in MySQL database, but can be hosted in any relational
database.
Fig. 1 Overall GDSS architecture within a BI framework
B. Architecture
The GDSS comprises of the Data, Application and Web
layers as described in the following:
1) Application layer hosts all the functional modules which
implement the methodology. The layers comprises of the
following major functional modules, which have been
developed in Java language:
? Group facilitation module. It is the core module of
GDSS, which is responsible for the coordination of
group decision process.
? Communication module. This module implements all
the necessary functionality, which is required for the
communication between group members and
Facilitator.
? Multicriteria algorithm module. This module
implements the NeXClass multicriteria classification
algorithm [17].
? Aggregation module. This module implements all the
aggregation processes following the methodology.
SJS and OWA aggregation are implemented on
individual members’ preferences [18], [19].
? Presentation module. This module is responsible for
the presentation formatting, in both simple numeric
and graphical formats. Utilization of graphs for result
visualization, increases familiarization and
understanding from group members. For the
development of graphs JChart library has been
utilized.
2) Data layer, stores all the necessary data for decision
problems. It is one of the core components of GDSS
architecture, and is responsible for storing all the
necessary data for each classification problem. Since the
orientation of GDSS is to operate into business
environment, the data model was designed to meet
relevant demands. It can store problem parameters from
multiple simultaneous decision problems and can handle
any combination of group members and parameters
without conflicts. It can also store previous problems or
demonstration ones for educational purposes, with
specific consideration to privacy issues. In order to meet
the above needs, we have implemented a relational model
distinguishing three major virtual groups of entities:
Problem parameters, Preferences and Results. Each one
consists of a number of Database tables which along with
the relations satisfy the needs of the GDSS.
? Problem parameters group stores all necessary data
related to a group problem. Parameters include all
necessary data for a decision problem referring to
criteria, categories, alternatives and members.
? Preferences group stores all the data representing
group members’ preferences. Preferences group is
separated into two sub groups, which store
individuals’ and aggregated preferences accordingly.
Aggregated preferences data is the input for the
multicriteria classification methodology.
? Results group stores all the data related to the results of
the problem.
? The model can reside in any relational database
available at business environment. In addition, there
is the option to import data in the form of XML
documents for decision problems with large number
of alternatives, when the data are already present into
another system. Data can be originated from several
sources of a firm’s BI infrastructure and further the
entire Data layer can be itself a part of the BI
infrastructure.
3) Web layer, provides all the user interface functionality.
User interface has been designed in order to guide users
through the steps of the methodology and has been
implemented using web technology. Servlets and html
pages offer GDSS functionality to group members in a
user friendly way. In addition, an XML interface has been
developed for importing data for large scale problems
which are already stored in existing systems. Finally, a
subset of GDSS functionality can be provided as a web
service.
GDSS is accessed through a main login page, where users
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 74
have to provide appropriate password. The system recognizes
two roles: Facilitator and Member. Facilitator works on a full
functional mode of the system, while group Members work on
a mode presenting a subset of functionality. Facilitator initiates
a new problem or selects to process an existing one. For a new
problem, he defines the proposed parameters and informs
group Members. For an existing problem, he can validate
Members’ input, and proceed to aggregation of preferences
and classification of the training set.
Members, after logging into the GDSS, can select a problem
and insert their preferences upon the proposed parameter set.
Several graphs provide visualizations over the numeric
parameters helping members’ understanding on them.
After validation of members’ input and aggregation of
preferences, facilitator executes the classification algorithm
using the appropriate functions from his menu, and informs
members for the classification results for the final assessment
phase (Fig. 2, Fig. 3, Fig.4).
Fig. 2 Facilitator’s mode for problem initiation
Fig. 3 Facilitator’s mode for category thresholds definition
Fig. 4 Member’s mode for parameter definition on criteria
V. CONCLUSION
In this paper we presented the prototype for a web based
Group Decision Support System for small collaborating teams
based on web technology highlighting the methodology and
the overall GDSS architecture and functionality. The GDSS is
based on web technology in order to be easily integrated
within existing business infrastructure. A layered approach was
followed, implementing Service Oriented Architecture and a
subset of the functionality of the GDSS is provided in terms of
Web Services. The system has been tested and evaluated
within banking environment, where it operates supporting
mainly financial decisions. Empirical findings from GDSS
application provide evidence that the methodology and the
GDSS provide a valid approach for similar decision problems
in business environment. We believe that this methodology
and GDSS can be easily deployed to support group decisions
in similar environments.
REFERENCES
[1] C. Hsinchun, R. HL Chiang, and V.C. Storey. "Business Intelligence
and Analytics: From Big Data to Big Impact", MIS quarterly vol. 36,
no. 4, 2012, pp. 1165-1188.
[2] A. McAfee, et al. "Big Data." The management revolution. Harvard
Bus Rev, vol. 90, no. 10, 2012, pp. 61-67.
[3] B. Mierzejewska, “Knowledge management”, Warsaw: Rzeczpospolita,
2002, Vol. 96.
[4] H.J. Watson, and B.H. Wixom, “The current state of business
intelligence”, Computer, vol40, 2007, pp. 96-99.
[5] C. M. Olszak, “The Purposes and Assumptions of Business
Intelligence.” In: Sroka, H. (Eds): Organizational Support Systems.
Katowice: Univerity of Economics. 2002.
[6] D. J. Power, “Decision Support Systems: From the Past to the Future.”,
Proceedings of the 2004 Americas Conference on Information Systems,
New York, NY, August 6-8, pp, 2025-2031.
[7] N. Karadimas, G. Rigopoulos, and A. Orsoni, “A Decision Model for
Group Assessment of Credit Applications”, in Proceedings of the 10th
International Conference on Computer Modelling & Simulation (IEEE),
pp. 319-323, Cambridge, England, April 2008.
[8] G. Rigopoulos, “Preference Aggregation for Group Decisions in
Management, International Scientific Conference”, eRA – 6, The
Conference for International Synergy in Energy, Environment, Tourism
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 75
and contribution of Information Technology in Science, Economy,
Society and Education, Athens, Greece, 2011.
[9] G. Rigopoulos, N. Karadimas, A. Orsoni, and J. Otamendi, “An
Interactive Web based Decision Support System for Financial Decisions
within a Team”, in Proceedings of the 4th International Conference on
Innovations in Information Technology (IEEE), pp. 546-550, November
2007, Dubai, United Arab Emirates
[10] Kimball, R. and M. Ross, (2002), The Data Warehouse Toolkit, 2nd ed.,
John Wiley and Sons, New York, NY.
[11] B. Liautaud, and M., Hammond, (2002), E-Business Intelligence.
Turning Information into Knowledge into Profit. New York: McGraw-
Hill, 2002.
[12] T. B. Pedersen, “How Is BI Used in Industry?: Report from a
Knowledge Exchange Network”, In DaWaK, 2004, pp. 179–188.
[13] T. B. Pedersen and C. S. Jensen, “Multidimensional Database
Technology”, IEEE Computer, 2001, vol. 34, no. 12, pp.40-46.
[14] R.T. Herschel, and N.E. Jones, “Knowledge management and business
intelligence: the importance of integration”, Journal of Knowledge
Management, Vol. 9, No 4, 2005, pp. 45-55.
[15] W. Cody, J. Kreulen, V. Krishna, and Spangler, W. “The integration of
business intelligence and knowledge management”. IBM Systems
Journal, Vol. 41, No. 4, 2002, pp. 697-713.
[16] H. Bhargava, and D. J. Power, “Decision Support Systems and Web
Technologies: A Status Report”, Proceedings of the 2001 Americas
Conference on Information Systems, Boston, MA, August 3 - 5, 2001.
[17] G. Rigopoulos, D. Askounis, and K. Metaxiotis, “NeXCLass: A
Decision Support System for non-ordered Multicriteria Classification”,
International Journal of Information Technology & Decision Making,
2010, vol. 9, no. 1, pp.53-79.
[18] G. Rigopoulos, J. Psarras, and D. Askounis, “An Aggregation Approach
for Group Multicriteria Assignment”, American Journal of Applied
Sciences vol. 5, no. 8, 2008, pp. 952-958.
[19] G. Rigopoulos, N. Karadimas, “Preference Aggregation for
Collaborative Multicriteria Assignment”, in Proceedings of the 1st
National joint Conference of Hellenic Mathematical Society and
Hellenic Operational Research Society, 2011, Athens, Greece.
George A. Rigopoulos. He holds a BSc in Physics and an MSc in Decision
Sciences with specialization in e-commerce. He also holds a PhD in MIS and
Operations research from National Technical University of Athens. His
research is oriented towards Management Information Systems, intelligent
decision support and multiagent simulation, as well as multicriteria decision
making in the business sector. He has wide working experience in IT projects,
decision support systems and e-business development. He has also wide
teaching experience in IT and e-business related subjects holding a visiting
lecturer position in Higher Technological Educational Institute of Athens.
Nikolaos V. Karadimas. Born in Athens and holds a
degree in Electrical Engineering from Technological
Institute of Patras. He holds a second degree in Electronic
Engineering and a Masters degree in "Computer Science"
from Glasgow Caledonian University, Scotland. He also
holds a second Masters degree in "Distributed and
Multimedia Information Systems" from Heriot-Watt
University, Edinburgh, Scotland. He holds a PhD in
Computer Engineering entitled "Resource Management and Decision Support
Systems Geo-Referred Information for a Logistics Environment" from the
National Technical University of Athens (NTUA). He teaches Informatics at
the Hellenic Military Academy since 2001 and since 2013 he holds a Lecturer
position in Informatics in the Hellenic Military Academy. He has also taught
at the Air Force Academy, at the Technological Educational Institute of
Piraeus, at the Technological Educational Institute of Chalkis, at the Hellenic
Air Force Technical NCO Academy and at the New York College. He has
worked at the Institute of Informatics and Telecommunications of National
Center for Scientific Research (Demokritos) and at the faculty of Electrical
and Computer Engineering of NTUA as an external collaborator - researcher.
In addition, he has supervised a large number of Diploma and Master’s theses
at the Hellenic Military Academy and at the Faculty of Electrical and
Computer Engineering of NTUA. He has obtained a number of scholarships
during his undergraduate and graduate studies and awards for four (4)
published articles. He is reviewer for many journals and conferences. Besides,
Dr. Karadimas has over sixty (60) publications in international journals and
conferences. He is a member of IEEE and IET and his research interests are in
the area of Military Applications, Databases, Resource Management,
Geographical Information Systems, Modeling and Simulation Algorithms and
Decision Support Systems.
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 76
doc_690483199.pdf
Collaborative Decisions within Business Intelligence Context A GDSS prototype
Abstract—In this work we present a prototype for a web based
Group Decision Support System, which can be integrated within a
firm’s Business Intelligence (BI) architecture to support decisions in
small collaborating teams. It is based on web technology and can be
used in asynchronous mode from group members. It implements a
multicriteria methodology for classification decisions where
aggregation of members’ preferences is performed at the parameter
level. Initially, a set of parameters defined by stakeholders is
proposed to the group by group facilitator. Next, each group member
evaluates the proposed parameter set and expresses her preferences in
numeric or linguistic format. Individual preferences are aggregated
by appropriate operators, and a group parameter set is produced
which is used as input for the classification algorithm. NeXClass
multicriteria classification algorithm is used for the classification of
alternatives, initially at a training set of alternatives and later at the
entire set. Finally, group members evaluate results, and consensus as
well as satisfaction metrics are calculated. In case of low acceptance
level, problem parameters are redefined by group members and
aggregation phase is repeated. The system has been utilized to solve
real world group classification problems, especially within the BI
environment of a commercial bank, supporting mainly financial
decisions. Empirical findings from the GDSS application have been
evaluated and enhancements have already been incorporated in order
to improve existing functionality and provide additional features.
Regarding the overall approach, findings provide evidence that it is
appropriate for similar decision problems in numerous business
environments, and the GDSS can be a valuable tool for enhancing a
BI framework with advanced decision capabilities.
Keywords—Business intelligence, Group decision support,
multicriteria decisions.
I. INTRODUCTION
OLLOWING advances in Information Technology the
majority of processes and decisions in large firms and
organizations can be characterized today as data driven. The
ability of acquisition and organization as well as development
and diffusion of knowledge has become a critical factor for
market performance and firm viability. Moreover, the amount
of information gathered from traditional and novel sources
such as customers, Internet and information systems, is
excessively increasing requiring additional organizational
effort. As a consequence, increasing complexity at the
G.A.R. is with the Higher Technological Educational Institute of Athens,
Athens, Greece (e-mail: [email protected]).
N.V.K. is with the Hellenic Military Academy, Faculty of Military
Sciences, Division of Mathematics and Engineering Sciences, Vari, Greece
(corresponding author’s contact details: +306932121552 e-mail:
[email protected]).
knowledge level has led to additional needs for advanced
decision support. Provision thus of appropriate decision
support tools at managerial as well as at operational level is
critical for efficient performance, and moreover it offers firms
a strong advantage against the rest of the competitors within
the domain [1], [2], [3].
One of the Information Technology directions that aims to
support firms handle the above complexity is Business
Intelligence (BI). BI provides a set of methods, processes and
tools to support firms’ decisions through intelligent
exploration, integration, aggregation and analysis of data from
various resources. In brief, the origins of BI tools can be
traced at the early developments of Decision Support Systems
(DSS), while current needs for advanced intelligent decisions
due to information complexity has led to a massive
development of BI systems with DSSs’ being a subset of BI
domain. Within a BI architecture, a DSS stands on top of BI
tools, utilizing aggregated information provided though them,
to assist decision makers optimize their decisions.
Additionally, since most of the decisions within firms today
require the participation of a group of decision makers, it is
critical to provide tools to assist them within the context of a
BI framework [4], [5], [6].
Group decision support has received significant attention
from researchers due to its potential application to various
business domains. Research in decision support systems
targets towards supplying decision makers with appropriate
tools to assist them in optimizing their decisions. Since a
decision support system has to reflect decision makers’
preferences or decision model, building an appropriate Group
Decision Support System (GDSS) is not a straightforward
process. Moreover, a number of issues have to be considered
such as individuals’ preference modeling, negotiation
protocols and coordination, to mention a few.
Several methodologies and tools have been developed in
order to support groups, ranging from collaborative techniques
to negotiation ones, depending on whether group members
share a common goal or support individual goals.
Technologies acquired for developing a GDSS tend to follow
advances in Information Technology, resulting in recent
advanced systems. Incorporation of web technologies
nowadays, for example, can support collaboration features
which could not be implemented in the very early GDSSs.
However, from literature analysis there are very few
approaches that combine BI concepts and Group Decision
Collaborative Decisions within Business
Intelligence Context: A GDSS prototype
George A. Rigopoulos, Nikolaos V. Karadimas
F
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 70
Support Systems (GDSS), which is the main contribution of
our work [7], [8], [9].
Following the above, we developed a structured
methodology which is based on multicriteria analysis and
supports group classification decisions and a GDSS which
implements it. In this paper we present the methodology and
architecture along with a prototype.
The overall architecture and development of the GDSS was
based on web technology in order to be easily integrated
within an existing BI infrastructure. We followed a layered
approach, implementing concepts from Service Oriented
Architecture (SOA), aiming to provide a subset of the
functionality of the GDSS in terms of Web Services. The
proposed GDSS can be part of an existing BI architecture
within a firm or an organization, gathering input from several
BI subsystems to integrate them into a decision support
framework.
The GDSS has been deployed in real business environment
supporting mainly financial decisions. Regarding the overall
methodology, findings provide evidence that it is a valid
approach for similar decision problems in numerous business
environments within a BI architecture.
The structure of the paper is as follows. Following the
introduction, relevant background information in BI and
decision support as well as group decisions is presented. Next,
we present the integrated group decision multicriteria
methodology implemented by the GDSS. Section 4 presents
the architecture of the developed GDSS mentioning key
features of the system. Finally, the conclusion summarizes the
work.
II. BACKGROUND AND RELEVANT WORK
A. BI and decision support
BI is generally defined as “a term to describe leveraging the
organization’s information assets for making better business
decisions” [10]. Intelligence in BI is often defined as the
discovery and explanation of hidden, inherent and decision-
relevant contexts in large amounts of business and economic
data. BI consists today one of the fastest developing domains
in Information Technology. It is widely assumed that in the
near future BI systems integration with CRM (Customer
Relationships Management) and ERP systems (Enterprise
Resource Planning) will provide firms a strong competitive
advantage, enhancing quality of managerial decisions [11], [5].
In general, BI systems combine data from internal
information systems of a firm and integrate with data coming
from the environment such as. statistics and financial
databases, to provide adequate and reliable up-to-date
information on different aspects of firms’ activities [5]. The
use of BI tools is popular in industry [12], [13], indicating the
firms’ growing needs to handle the vast aggregation of
information orienting from business documents and data,
including business forms, databases, spreadsheets, e-mails,
articles, technical journals and reports, contracts, and web
documents. Distinction between knowledge management and
BI is not always clear [14], although knowledge and content
management technologies search, organize and extract value
from information sources, while BI focuses on the same
purpose, but from a different scope.
BI is mainly targeting in advancing decision making
utilizing data warehousing and online analytical processing
techniques (OLAP), collecting relevant data into repositories,
where organized and validated can be further available for
decision making. In general, business data are extracted,
transformed and loaded from various transactional systems
into a data warehouse after data cleansing processes and
multidimensional models are created to support drill down and
roll up analyses. A number of vendors provide tools and
platforms for such operations and advanced end user
functionality to support large amount of data [15].
From the above, the linkage between BI and decision
support within firms is evident. Moreover, BI origins can be
traced back at the early data-driven DSS approaches [6]. Later,
BI term was promoted and used as an umbrella term “to
describe a set of concepts and methods to improve business
decision making by using fact-based support systems”.
Although BI is sometimes used instead of the term of
executive information systems (EIS), in general BI systems
can be defined as data-driven DSSs. With the advent of
Internet, BI vendors shifted their BI solutions towards web
technologies and enterprise BI portals emerged [16].
B. Group decisions
Group decision making has become an essential component
of both strategic planning and everyday operations for the
majority of today’s organizations and enterprises. Since
complexity of business environment requires sufficient
knowledge from a wide range of domains, contribution of a
team of experts with key skills is the only way to achieve
efficiency in decisions. In order to support groups’ needs,
various researchers work on developing tools and
methodologies, ranging from collaboration technologies to
decision support systems [7], [8].
However, group decisions are quite more complex
compared to single decision making, since a number of
contradicting factors are involved such as individuals’ personal
opinions, goals and stakes resulting in a social procedure,
where negotiation and strategy plays a critical role.
Group decision making in real business environments raises
also some issues such as:
• Conflicting individual goals,
• Not efficient knowledge,
• Validity of information,
• Individuals’ motivation.
Despite the inherent complexity, within a group decision
making setting, a member is able to express personal opinions
and suggest solutions from a personal perspective. In addition,
negotiation and voting advance efficiency of decisions and
increment acceptability and adoption since all participants
have contributed to the result, smoothening thus any disputes.
In general, group members can be motivated by individual
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 71
perceptions to work within the group either towards
collaboration or towards competition. While in the first case,
members express similar opinions and goals, in the second one
they state opposite opinions. Although collaborative teams
work towards a common goal contradictions may also occur.
Some key techniques that have been acquired in order to
facilitate group work and decision include:
• Brainstorming,
• Nominal group technique,
• Delphi method,
• Voting,
• Multicriteria analysis.
C. Group decisions and multicriteria analysis
Group decision support is a subset of the more extended
research area of group support and negotiation. Group
decision support is an active research topic and existing
literature is quite extensive covering business as well as social
issues. Limiting the scope of relevant literature to integration
of multicriteria analysis within group decisions, we performed
a detailed survey focusing on relevant approaches, and
especially on developed Group decision support systems. An
extensive review of multicriteria analysis integration within
GDSSs is presented by Rigopoulos [7], [8], [9].
In general, multicriteria analysis can be incorporated as a
method to model preferences and facilitate decision making
within a group of decision makers. Modeling under a
multicriteria setting can be formulated under two major
directions:
1) In the first approach, individual multicriteria models are
developed, which capture individuals’ preferences. Each
group member formulates a multicriteria problem defining
the parameters according to her preferences and solves the
problem getting an individual solution set. Next, the
separate solutions are aggregated by aggregation operators
providing thus the group solution.
2) In the second approach, one multicriteria model is
developed for the entire team. Each group member
provides a set of parameters which are aggregated by
appropriate operators, providing finally a group parameter
set. Upon this set the muticriteria method is applied and
the solution expresses the group preference.
Each approach poses both positive and negative aspects
depending on the aggregation operation, which is followed.
III. PROPOSED GROUP DECISION METHODOLOGY
The main objective of our work is to provide support to a
group of decision makers in classification problems. The
problem refers to the assignment of a set of alternatives in a
number of predefined non-ordered categories, according to
their performance on a set of evaluation criteria. For this
reason, we have developed a structured group decision
methodology, which is based on the following principles:
1) The decision group is a small homogeneous team of
collaborating decision makers. Although the methodology
can be extended to large decision teams, our approach is
based on collaborative teams, which target towards
maximizing consensus. Non-collaborative teams require a
negotiation-based approach, which is out of scope of the
present methodology.
2) A facilitator coordinates the entire decision process. The
entire group decision process is coordinated by a
Facilitator. Usually, in group decision making a
negotiation phase takes place at the preliminary steps of
the decision problem formation. During this negotiation,
which can be structured or not, basic parameters are
defined. Since our methodology is not focusing on group
formation procedure and initial negotiations, we consider
that a preliminary negotiation step has already been
executed, possibly by utilizing brainstorming technique,
between stakeholders, and the outcome of this process is
an initial proposal of parameters. This set is expressed
from Facilitator as the initial proposal upon which group
members will express their preferences. Facilitator guides
the entire process in order to produce efficient and timely
results.
3) Decision problem is structured or semi structured. The
team solves a structured classification problem
contributing their preferences. Non structured problems
are out of scope.
4) Multicriteria analysis is utilized for the classification. For
the classification problem we utilize multicriteria analysis
which provides appropriate support to similar problems.
Following the above principles, we developed a group
decision methodology which is separated in the following
major phases:
A. Problem initiation
. In this phase Facilitator initiates the decision problem,
defining all appropriate parameters. The parameters are related
to the specific multicriteria methodology, and refer to criteria,
alternatives and categories. In details:
1) Basic parameters. Initially, Facilitator defines a number of
basic parameters, related to classification problem such as
the number of group members, the number of categories,
the number of criteria, and to results assessment such as
the consensus, satisfaction and acceptance levels. These
levels define minimum required levels for the group
decision. In case they are not satisfied, a second round is
executed with modification of individual preferences.
2) Members. Facilitator defines the group members assigning
all necessary contact details.
3) Categories. Facilitator defines the set of categories for the
classification of alternatives.
4) Evaluation criteria. Facilitator defines the set of
evaluation criteria according to problem requirements.
5) Criteria weights. Facilitator defines the criteria weights.
6) Alternatives. Facilitator defines the set of alternatives for
classification, and defines their performance on the
evaluation criteria.
7) Entrance thresholds. Facilitator defines appropriate
entrance thresholds for each category.
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 72
8) For each threshold Facilitator defines preference,
indifference and veto thresholds similar to ELECTRE TRI
method.
9) Training set. Facilitator defines a subset of alternatives as
training set, in order to evaluate parameters’ accuracy.
After the initiation of the parameters, Facilitator
communicates through the GDSS with group members
informing them about the problem and asking them to submit
their preferences.
B. Aggregation of individual parameters
In this phase group members express their preferences on
the proposed parameter set. Member preferences are expressed
in numeric values and linguistic preferences. For the
aggregation of numeric values we utilize the Social Judgment
Scheme (SJS), while linguistic terms are aggregated in terms
of an Ordered Weighted Averaging Operator (OWA) [18],
[19].
Aggregation of member preferences is executed for the
following parameters
1) Criteria. Group members express their acceptance on each
proposed criterion in a five point linguistic scale and their
preferred weight in numeric value.
2) Alternatives. Group members express their acceptance on
alternatives’ performance or submit their preference in
numeric value.
3) Categories. Group members express their acceptance on
each category definition, and submit their preferences on
category thresholds in numeric value. .
Facilitator proceeds with validation of members’ input and
aggregates the values. Parameters with low acceptance level
are marked and are subject to review if final results are not
acceptable by group members.
C. Application of multicriteria classification algorithm
After the aggregation of individual members’ parameters a
group parameter set is created and NeXClass algorithm for
multicriteria classification is applied on this group parameter
set [17]. NeXClass algorithm classifies an alternative to a
specific category with respect to alternative’s performance to
the evaluation criteria, considering a set of alternatives, a set of
predefined non-ordered categories and a set of evaluation
criteria. In more details the algorithm works as follows:
1) For each category decision maker defines an entrance
threshold using available information. This threshold
represents the minimum requirements for an alternative in
terms of performance on the evaluation criteria in order to
be included in this category.
2) Decision maker defines alternatives’ performance on the
evaluation criteria.
3) For each alternative an excluding degree is calculated for
every category threshold, based on outranking relations,
following a similar approach to ELECTRE TRI method.
4) Next, the fuzzy excluding degree of an alternative over a
category is calculated.
5) Assignment to a category is based on the rule which states
that alternative is assigned to the category for which the
excluding degree over the entrance threshold is minimum.
Application of NeXClass classification algorithm is
executed through the following steps
1) Training set classification. Classification algorithm is
initially applied to the training set initially, as it has been
defined by group members. Classification is executed by
Facilitator, and group members are informed to assess the
results.
2) Evaluation of results. Each member expresses her
preference on the results in a five point linguistic scale,
and in case of low acceptance level, Facilitaror executes a
second round of parameter definition from members in
order to calibrate the model. When training set
classification is acceptable, Facilitator proceeds with the
classification of the entire set of alternatives. In case of
low acceptance level after the second round, Facilitator
terminates the process in order to revise the problem with
stakeholders.
3) Training set classification. Classification algorithm is
finally applied to the entire set by Facilitator, and group
members are informed to assess the results.
D. Results assessment
Group members assess the results expressing their
preference in a five point linguistic scale. In case of low
acceptance level, Facilitator reruns the model, requesting
modifications from members.
IV. GROUP DECISION SUPPORT SYSTEM
A. Overview
A prototype GDSS was developed to implement the
aforementioned methodology. The design of the GDSS was
based on the following requirements:
1) Collaboration. The GDSS should promote collaboration
between group members by appropriate functions. Group
members for similar decision problems, can be selected ad
hoc without any prior collaboration. The GDSS should
thus promote the feeling of a common goal to members
minimizing thus individual goals.
2) Communication. Since business operations may span over
several locations, members can be located separately.
Communication thus between facilitator and group
members should be efficient enough in order to provide
results in a timely way. The GDSS should thus provide
appropriate communication tools.
3) Anonymity. Although anonymity poses some negative
issues, it encourages members express their preferences
without restrictions or external influence. For this reason,
the GDSS should support anonymity at presentation level.
4) Asynchronous operation. Different time zones and
locations of today’s business operations require
asynchronous operation and decision making. The GDSS
should thus provide asynchronous operation efficiently.
Considering the above requirements, a layered approach
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 73
which can be easily deployed in an existing BI infrastructure
was selected for the GDSS architecture (Fig. 1). BI
architecture components such as transactional data source
systems and data warehouses can be easily connected and
integrated with the GDSS, being the sources of input data. In
the same way, GDSS output can be deployed in firm’s BI
systems advancing firm’s knowledge. Regarding the
technology utilized, GDSS modules have been entirely
developed in Java language using JCharts library for chart
preparation. Apache web server is used to host the entire site
with Tomcat as servlet container and Tomcat Axis used for the
deployment of web services. Data layer has been implemented
in MySQL database, but can be hosted in any relational
database.
Fig. 1 Overall GDSS architecture within a BI framework
B. Architecture
The GDSS comprises of the Data, Application and Web
layers as described in the following:
1) Application layer hosts all the functional modules which
implement the methodology. The layers comprises of the
following major functional modules, which have been
developed in Java language:
? Group facilitation module. It is the core module of
GDSS, which is responsible for the coordination of
group decision process.
? Communication module. This module implements all
the necessary functionality, which is required for the
communication between group members and
Facilitator.
? Multicriteria algorithm module. This module
implements the NeXClass multicriteria classification
algorithm [17].
? Aggregation module. This module implements all the
aggregation processes following the methodology.
SJS and OWA aggregation are implemented on
individual members’ preferences [18], [19].
? Presentation module. This module is responsible for
the presentation formatting, in both simple numeric
and graphical formats. Utilization of graphs for result
visualization, increases familiarization and
understanding from group members. For the
development of graphs JChart library has been
utilized.
2) Data layer, stores all the necessary data for decision
problems. It is one of the core components of GDSS
architecture, and is responsible for storing all the
necessary data for each classification problem. Since the
orientation of GDSS is to operate into business
environment, the data model was designed to meet
relevant demands. It can store problem parameters from
multiple simultaneous decision problems and can handle
any combination of group members and parameters
without conflicts. It can also store previous problems or
demonstration ones for educational purposes, with
specific consideration to privacy issues. In order to meet
the above needs, we have implemented a relational model
distinguishing three major virtual groups of entities:
Problem parameters, Preferences and Results. Each one
consists of a number of Database tables which along with
the relations satisfy the needs of the GDSS.
? Problem parameters group stores all necessary data
related to a group problem. Parameters include all
necessary data for a decision problem referring to
criteria, categories, alternatives and members.
? Preferences group stores all the data representing
group members’ preferences. Preferences group is
separated into two sub groups, which store
individuals’ and aggregated preferences accordingly.
Aggregated preferences data is the input for the
multicriteria classification methodology.
? Results group stores all the data related to the results of
the problem.
? The model can reside in any relational database
available at business environment. In addition, there
is the option to import data in the form of XML
documents for decision problems with large number
of alternatives, when the data are already present into
another system. Data can be originated from several
sources of a firm’s BI infrastructure and further the
entire Data layer can be itself a part of the BI
infrastructure.
3) Web layer, provides all the user interface functionality.
User interface has been designed in order to guide users
through the steps of the methodology and has been
implemented using web technology. Servlets and html
pages offer GDSS functionality to group members in a
user friendly way. In addition, an XML interface has been
developed for importing data for large scale problems
which are already stored in existing systems. Finally, a
subset of GDSS functionality can be provided as a web
service.
GDSS is accessed through a main login page, where users
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 74
have to provide appropriate password. The system recognizes
two roles: Facilitator and Member. Facilitator works on a full
functional mode of the system, while group Members work on
a mode presenting a subset of functionality. Facilitator initiates
a new problem or selects to process an existing one. For a new
problem, he defines the proposed parameters and informs
group Members. For an existing problem, he can validate
Members’ input, and proceed to aggregation of preferences
and classification of the training set.
Members, after logging into the GDSS, can select a problem
and insert their preferences upon the proposed parameter set.
Several graphs provide visualizations over the numeric
parameters helping members’ understanding on them.
After validation of members’ input and aggregation of
preferences, facilitator executes the classification algorithm
using the appropriate functions from his menu, and informs
members for the classification results for the final assessment
phase (Fig. 2, Fig. 3, Fig.4).
Fig. 2 Facilitator’s mode for problem initiation
Fig. 3 Facilitator’s mode for category thresholds definition
Fig. 4 Member’s mode for parameter definition on criteria
V. CONCLUSION
In this paper we presented the prototype for a web based
Group Decision Support System for small collaborating teams
based on web technology highlighting the methodology and
the overall GDSS architecture and functionality. The GDSS is
based on web technology in order to be easily integrated
within existing business infrastructure. A layered approach was
followed, implementing Service Oriented Architecture and a
subset of the functionality of the GDSS is provided in terms of
Web Services. The system has been tested and evaluated
within banking environment, where it operates supporting
mainly financial decisions. Empirical findings from GDSS
application provide evidence that the methodology and the
GDSS provide a valid approach for similar decision problems
in business environment. We believe that this methodology
and GDSS can be easily deployed to support group decisions
in similar environments.
REFERENCES
[1] C. Hsinchun, R. HL Chiang, and V.C. Storey. "Business Intelligence
and Analytics: From Big Data to Big Impact", MIS quarterly vol. 36,
no. 4, 2012, pp. 1165-1188.
[2] A. McAfee, et al. "Big Data." The management revolution. Harvard
Bus Rev, vol. 90, no. 10, 2012, pp. 61-67.
[3] B. Mierzejewska, “Knowledge management”, Warsaw: Rzeczpospolita,
2002, Vol. 96.
[4] H.J. Watson, and B.H. Wixom, “The current state of business
intelligence”, Computer, vol40, 2007, pp. 96-99.
[5] C. M. Olszak, “The Purposes and Assumptions of Business
Intelligence.” In: Sroka, H. (Eds): Organizational Support Systems.
Katowice: Univerity of Economics. 2002.
[6] D. J. Power, “Decision Support Systems: From the Past to the Future.”,
Proceedings of the 2004 Americas Conference on Information Systems,
New York, NY, August 6-8, pp, 2025-2031.
[7] N. Karadimas, G. Rigopoulos, and A. Orsoni, “A Decision Model for
Group Assessment of Credit Applications”, in Proceedings of the 10th
International Conference on Computer Modelling & Simulation (IEEE),
pp. 319-323, Cambridge, England, April 2008.
[8] G. Rigopoulos, “Preference Aggregation for Group Decisions in
Management, International Scientific Conference”, eRA – 6, The
Conference for International Synergy in Energy, Environment, Tourism
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 75
and contribution of Information Technology in Science, Economy,
Society and Education, Athens, Greece, 2011.
[9] G. Rigopoulos, N. Karadimas, A. Orsoni, and J. Otamendi, “An
Interactive Web based Decision Support System for Financial Decisions
within a Team”, in Proceedings of the 4th International Conference on
Innovations in Information Technology (IEEE), pp. 546-550, November
2007, Dubai, United Arab Emirates
[10] Kimball, R. and M. Ross, (2002), The Data Warehouse Toolkit, 2nd ed.,
John Wiley and Sons, New York, NY.
[11] B. Liautaud, and M., Hammond, (2002), E-Business Intelligence.
Turning Information into Knowledge into Profit. New York: McGraw-
Hill, 2002.
[12] T. B. Pedersen, “How Is BI Used in Industry?: Report from a
Knowledge Exchange Network”, In DaWaK, 2004, pp. 179–188.
[13] T. B. Pedersen and C. S. Jensen, “Multidimensional Database
Technology”, IEEE Computer, 2001, vol. 34, no. 12, pp.40-46.
[14] R.T. Herschel, and N.E. Jones, “Knowledge management and business
intelligence: the importance of integration”, Journal of Knowledge
Management, Vol. 9, No 4, 2005, pp. 45-55.
[15] W. Cody, J. Kreulen, V. Krishna, and Spangler, W. “The integration of
business intelligence and knowledge management”. IBM Systems
Journal, Vol. 41, No. 4, 2002, pp. 697-713.
[16] H. Bhargava, and D. J. Power, “Decision Support Systems and Web
Technologies: A Status Report”, Proceedings of the 2001 Americas
Conference on Information Systems, Boston, MA, August 3 - 5, 2001.
[17] G. Rigopoulos, D. Askounis, and K. Metaxiotis, “NeXCLass: A
Decision Support System for non-ordered Multicriteria Classification”,
International Journal of Information Technology & Decision Making,
2010, vol. 9, no. 1, pp.53-79.
[18] G. Rigopoulos, J. Psarras, and D. Askounis, “An Aggregation Approach
for Group Multicriteria Assignment”, American Journal of Applied
Sciences vol. 5, no. 8, 2008, pp. 952-958.
[19] G. Rigopoulos, N. Karadimas, “Preference Aggregation for
Collaborative Multicriteria Assignment”, in Proceedings of the 1st
National joint Conference of Hellenic Mathematical Society and
Hellenic Operational Research Society, 2011, Athens, Greece.
George A. Rigopoulos. He holds a BSc in Physics and an MSc in Decision
Sciences with specialization in e-commerce. He also holds a PhD in MIS and
Operations research from National Technical University of Athens. His
research is oriented towards Management Information Systems, intelligent
decision support and multiagent simulation, as well as multicriteria decision
making in the business sector. He has wide working experience in IT projects,
decision support systems and e-business development. He has also wide
teaching experience in IT and e-business related subjects holding a visiting
lecturer position in Higher Technological Educational Institute of Athens.
Nikolaos V. Karadimas. Born in Athens and holds a
degree in Electrical Engineering from Technological
Institute of Patras. He holds a second degree in Electronic
Engineering and a Masters degree in "Computer Science"
from Glasgow Caledonian University, Scotland. He also
holds a second Masters degree in "Distributed and
Multimedia Information Systems" from Heriot-Watt
University, Edinburgh, Scotland. He holds a PhD in
Computer Engineering entitled "Resource Management and Decision Support
Systems Geo-Referred Information for a Logistics Environment" from the
National Technical University of Athens (NTUA). He teaches Informatics at
the Hellenic Military Academy since 2001 and since 2013 he holds a Lecturer
position in Informatics in the Hellenic Military Academy. He has also taught
at the Air Force Academy, at the Technological Educational Institute of
Piraeus, at the Technological Educational Institute of Chalkis, at the Hellenic
Air Force Technical NCO Academy and at the New York College. He has
worked at the Institute of Informatics and Telecommunications of National
Center for Scientific Research (Demokritos) and at the faculty of Electrical
and Computer Engineering of NTUA as an external collaborator - researcher.
In addition, he has supervised a large number of Diploma and Master’s theses
at the Hellenic Military Academy and at the Faculty of Electrical and
Computer Engineering of NTUA. He has obtained a number of scholarships
during his undergraduate and graduate studies and awards for four (4)
published articles. He is reviewer for many journals and conferences. Besides,
Dr. Karadimas has over sixty (60) publications in international journals and
conferences. He is a member of IEEE and IET and his research interests are in
the area of Military Applications, Databases, Resource Management,
Geographical Information Systems, Modeling and Simulation Algorithms and
Decision Support Systems.
Recent Advances in Electrical Engineering and Educational Technologies
ISBN: 978-1-61804-254-5 76
doc_690483199.pdf