CC BY
95
ISSN 1992-6502 (Print)_TB&C/ltVH/Ü/IC ^¡jf/jQ^^lj _ISSN 2225-2789 (Online)
Vol. 18, no. 5 (66), pp. 3-7, 2014 http://journal.ugatu.ac.ru
UDC 004.7
A SURVEY ON APPLICATIONS OF FUZZY COGNITIVE MAPS IN BUSINESS AND MANAGEMENT
1 2 O. D. Ntarlas , P. P. Groumpos
1 ntarlas@ece.upatras.gr, 2 groumpos@ece.upatras.gr University of Patras, Patras, Greece Submitted 2014, June 10
Abstract. The aim of this article is to present a survey on applications of fuzzy cognitive maps to today's business management indicating its comparative advantages to conventional expert systems. Thus, a survey on FCM studies related to business and management is elaborated, marking down their evolution during the last decade. Various representative applications of fuzzy cognitive maps are illustrated in diverse areas of business management, such as e business, performance assessment, human resources management, planning and investment decision making processes.
Keywords: fuzzy cognitive maps; e business; performance assessment; management.
1. INTRODUCTION
Nowadays, due to constant change of business conditions, adaptability is considered a significant competitive advantage for an organization. To achieve this, practical problem solving in real time is necessary [1]. For the last 40-50 years all problems originated from business processes needed experts, in order to be solved in a realistic and cost-effective way. Unfortunately, practical problems can arise at any time and experts cannot always be available. However, their knowledge could be utilized at all times in certain problematic areas with the use of a well-designed expert system (ES).
ES constitute the most commonly applied branch of Artificial Intelligence (AI). An ES is in fact a computational program, which represents and analyzes an expert's knowledge in a specific area, so as to advise or solve problems in that field. It emulates human reasoning regarding a problem by performing a complex process for decision making in a problem domain. Expert systems, since their development have been providing us with effective solutions-answers in a variety of problems-questions that are difficult to handle using other traditional methodologies.
Various ES methodologies [2] have led expert systems to overcome their limitations regarding their rule-based nature. However, there is still a long way to go. ES development continues to be a time-consuming and expensive task. Additionally, they can solve complex problems in a very narrow area and it is not always easy for experts to communicate their domain-specific knowledge. ES are
often mistrusted by their users, i.e. managers, who cannot get used to the idea that a computer program is going to decide instead of them. Managers seem to be fonder of tools for decision facilitation rather than automatic decision making.
It is a common truth that the success of business management lies in the consideration and provision of how a variety of factors interact between them. Today, the number of factors that must be taken into account for an effective business management has increased significantly, due to the highly dynamic business environment. Unfortunately, many times managers lack the ability of evaluating all the related factors, as they use to analyze and assess the impact of two to three factors simultaneously at best. It is very common for a business system to contain uncertain and fuzzy knowledge, due to the fact that most knowledge is expressed in terms of cause and effect. In addition, every business practitioner, i.e. expert, tends to have its individual point of view about effective business management. Fuzzy Cognitive Maps (FCMs) have come to fill that gap, as they are best suited for problems where experts have different opinions about a correct answer and they have the ability to model uncertain and fuzzy knowledge [51].
FCMs have comprised a tool used for decision facilitation in diverse scientific areas, such as medicine [67] and political science [68]. The purpose of this study is to review recent applications of Fuzzy Cognitive Maps in the domain of business. By doing that, it can be shown how FCMs can make life for managers a lot easier and it can be derived that
FCMs can constitute a useful tool for decision support in business management, too.
This paper has been organized in the following way. Section II gives a brief overview of Fuzzy Cognitive Maps, while Section III presents recent applications of them to various areas of business management. In Section IV a discussion on survey results takes place and the significant increase of FCM-related studies in business and management during the last decade is demonstrated. Finally, in Section V, some concluding remarks on Fuzzy Cognitive Mapping Technique are made.
2. A BRIEF OVERVIEW OF FUZZY COGNITIVE MAPS
Fuzzy Cognitive Maps (FCMs) is a modeling technique, arising from the combination of Fuzzy Logic and Neural Networks. Compared to conventional expert systems, FCMs have considerable advantages; they are dynamic feedback systems [5] and they can represent structured knowledge very easily, providing the higher level of abstraction needed in many cases [6]. FCMs constitute a useful tool, with which we take advantage and quantify the accumulated knowledge obtained through years of observing the operation and behavior of complex systems [7].
The term of "fuzzy cognitive map" was first used by Kosko [8] to illustrate a graphically signed directed graph with feedback, consisting of nodes and weighed arcs. The FCM's nodes represent the concepts used for describing system behavior. These nodes are interconnected by signed and weighted arcs, standing for the existing causal relationships between the concepts. In other words, the arcs describe the variation on the value of one concept when the value of an interconnected concept is being altered.
When concept C. influences concept C¿, there
is a w arc which can take any value between -1 and 1, quantifying this way the correlation between the two concepts. wjt can be:
• Positive (w > 0) when there is a positive causality between C ■ and C, i.e. an increase / decrease in the value of C brings an increase/decrease in the value of C respectively.
• Negative (wJt < 0) when the causality is negative and increase/decrease of the value of C causes a decrease/increase of the value of C .
• Zero (w = 0) when there is no influence of concept C. to concept C.
The bigger the absolute value of wp it is, the stronger the influence of C. to C will be, in a
direct (positive causality) or an inverse way (negative causality).
In a conventional FCM, the value of each concept is computed, taking into account the influence of other concepts to the specific concept [9], by applying the following mathematical procedure:
N
4(i+1) = fk ■ 4m + k-Zj ■ w,). (1)
j *'
j=i
In equation (1), A is the value of concept
C at time k +1,A. is the value of concept C,
1 j<k) r i
at time k , w is the weight of interconnection between concepts C and C and f is the sigmoid threshold function.
Fig 1. A Simple Fuzzy Cognitive Map
3. APPLICATIONS IN BUSINESS AND MANAGEMENT
Considering the abovementioned favorable features of the FCMs, besides the overall increase of FCM-related studies [10], we have observed that there has been also an increasing amount of literature on business and systems management in the recent years. In this work, we attempt to mark down FCM studies related to business and management and illustrate some interesting applications of them.
Several studies have succeeded in measuring and assessing business performance. Chytas et al. [11] managed to develop a new proactive balanced scorehand methodology with the aid of FCMs, which were used for quantifying the imprecise causal interrelationships of key performance indicators, so as to adjust performance goals. Of
course, it is not the only study which makes use of FCMs, in order to model and analyze business performance metrics [12]. In another work, FCMs were implemented in strategy maps, eliminating their drawbacks and providing them with competitive advantages in terms of flexibility and simulated scenarios. Results of case studies showed that the developed system could emulate effectively experts' predictions [13]. FCMs are used as a performance modeling tool for the implementation of business performance re-engineering methodology, in order to simulate the operational efficiency of complex, imprecise functional relationship and quantify the impact of BPR activities to the business model [14].
FCMs have found a great applicability in the planning process, too. Lee et al. [15] employed FCMs to industrial marketing planning. By integrating agent and FCM technology, they managed to overcome the conventional FCMs' limitations in marketing planning. Hence, experts' opinions from different functional departments were integrated and complex, ambiguous causalities among the related variables were quantified, allowing this way a systematic what-if analysis to be carried out, in order to compare various scenarios. Comparison and evaluation of different scenarios is done in another work too [16], in which FCMs are applied to enterprise resource planning, modeling maintenance risks. With the help of FCMs, ERP practitioners highlight the most important factors and are able to handle the maintenance risks more effectively. Kardaras and Karakostas [17] used FCMs as a supplement to the strategic planning of information systems. This way, it could be much easier for planners to label specific IT projects and evaluate their impact on an organization. Regarding business informatics, an interesting work was that, which proposed augmented FCMs for modeling LMS critical success factors, in order to discern the necessary activities for success [18].
FCMs have also addressed the growing need to assess investment decision-making processes. Irani et al. [19] managed to shed light upon the often cloudy evaluation of IS investments by identifying, classifying and correlating the factors that affected more the IS investment evaluation with the related knowledge components. As a result, an analysis of knowledge-based decisions in the IS evaluation investment area took place. In 2004, Koulouriotis [20] assigned FCMs to emulate investors' reasoning process, as a means for stock prediction.
Human Resources Management has also been implemented with the fuzzy cognitive mapping technique. Xirogiannis et al. [21] attempted to
model the operational efficiency of HRM activities and their impact to shareholder satisfaction. Thus, the effects of HR practices to the overall shareholder value were clearly illustrated. In another work, Motlagh [22] investigated HRM influence to Six Sigma projects implementation. The critical factors were identified, so managers knew where should focus, in order to achieve better organizational performance as far as HRM is concerned. Last but not least, FCM applications, such as modeling core value systems [23] and relationship management [24] enable possible detection of conflicts among colleagues, doing in this manner, the work of HRM much easier.
As far as e-business is concerned, Miao et al. [25] integrated Intelligent Agents into Fuzzy Cognitive Mapping, creating the tool of Fuzzy Cognitive Agents, in order to support e-commerce applications. That new approach enabled personalized recommendations based on the online user's preferences, common preferences of other users in general and expert's knowledge. Hence, they addressed the difficulty of many users, who cannot determine what they really want, when visiting electronic commerce web-sites. Also, Lee and Ahn [6] combined FCMs and structural equation modeling for control design support of an e-commerce web based system (ECWS), in order to achieve high ECWS performance. The FCM's fuzzy edges facilitated the representation of environmental factors, controls, implementation, and performance in ECWS, while the structural equation modeling defined the complex causal relationships among factors. In another publication, Xirogiannis and Glykas [26] used FCMs as a means for a causal representation of dynamic e-business maturity indicators. They showed how FCMs can work as a supplement to strategic change projects, by modeling complex strategic models and quantifying the impact of strategic changes to the overall e-business efficiency.
Other worth to be mentioned applications follow. Nasserzadeh et al. [27] used FCM modeling, in order to assess customer satisfaction, a competitive advantage in today's societies, in the sector of banking industry. The Delphi methodology, which included feedback reports for everyone, was exploited for easier consensus among the experts. The resulting FCM was capable of evaluating a bank's capacities for attracting new customers and increase customer satisfaction rate. Irani et al. [28] used fuzzy cognitive mapping to investigate the relationship between knowledge management and organizational learning. They showed that a relationship between them did exist and the constructed FCM helped them to identify
factors, with which a company could evolve to a learning organization. Wang et al. [29] used structural equation model as a supplement to FCM, in order to define the complex causal relationships among factors in a hotel environment. The constructed FCM enlightened the often cloudy context of building competitive advantages in the hotel industry and it enabled a systematic what-if analysis to be carried out, in order to compare various possible scenarios in reality. FCMs were applied in the complex sector of retail industry, too. They led to a better implementation of collaborative planning, forecasting and replenishment (CPFR) approach, as they highlighted the related factors for CPFR success [30].
4. DISCUSSION ON SURVEY RESULTS
In reviewing the literature, it has been observed that FCMs, have strong and weak points. They are flexible, adaptable and they can model very well semi-structured or unstructured knowledge. Nevertheless, they contribute to the problem solving process only by evaluating the alternatives of a scenario, thus, not absolving the manager from making the actual decision. The preceding features suggest that FCMs cannot be utilized in all kinds of problems. They are preferred in unstructured cases, where they can provide managers with very good results without the ethical implication of human replacement.
Table
Recent FCM Studies in Business & Management
Number of FCM-related studies in Business &
Year Management
Studies Journals & Book Chapters Conferences
2004 8 5 3
2005 1 0 1
2006 5 2 3
2007 7 5 2
2008 9 3 6
2009 12 5 7
2010 20 16 4
2011 20 11 9
2012 15 12 3
2013 22 15 7
Through an extensive survey and classification of FCM published studies across the internet, last decade's FCM applications in business and management are marked down in Table. There has been a steady increase in the number of FCM-related studies in this field, as it can be seen in Table. In particular, FCM studies regarding Business & Management in 2013 were almost tripled compared with those ten years ago. Considering the above-
mentioned, it is clear that FCM application in business and management has gained a considerable interest over the last ten years.
5. CONCLUSIONS
Nowadays successful decision making usually requires complicated mathematical models. Business problems are no exception to this rule, as they are characterized by controversial theories and mathematical solutions, if any, with complex equations and formulas. These difficulties have led to an extensive use of expert systems to problems of this domain, thanks to their capability of taking advantage of experts' domain-specific knowledge and emulate their inference in a very narrow area. This study set out to demonstrate various applications of Fuzzy Cognitive Maps to business management.
FCMs have been proven through literature, as a very useful cognition tool to identify and assess the key-factors of a variety of business issues. However, FCM technique cannot act as a panacea for solving business problems as it depends upon several assumptions, such as the existence of only symmetric or monotonic causal relationships. A lot of effort has been made for conventional FCMs to overcome their drawbacks either by using various supplements, or by advancing FCM theory and learning. Despite their drawbacks though, the increasing number of FCMs' applications not only in various aspects of business, but also in diverse scientific fields, seems quite promising. Hence, more research on FCMs is expected in the future, so as to evolve even more, leaving their limitations behind.
REFERENCES
1. M. Dasic, S. Trajkovic and B. Tesanovic, "The necessity of using expert systems in strategic decision making," Int. J. of Economics & Law, vol. 1, no. 1, pp. 27-35, Apr. 2011.
2. S. H. Liao, "Expert system methodologies and applications—a decade review from 1995 to 2004," Expert Systems with Applications, vol. 28, issue 1, pp. 93-103, Jan. 2005.
3. A. P. Anninou, P. P. Groumpos, P. Polychronopoulos, "Modeling health diseases using competitive fuzzy cognitive maps," Artificial Intelligence Applications and Innovations, pp. 88-95, Springer Berlin Heidelberg, 2013.
4. A. K. Tsadiras, I. Kouskouvelis, K. G. Margaritis, "Using fuzzy cognitive maps as a decision support system for political decisions," Advances in Informatics, pp. 172-182, Springer Berlin Heidelberg, 2003.
5. W. R. Taber, "Knowledge processing with fuzzy cognitive maps," Expert Systems with Applications, vol. 2, issue 1, pp. 83-87, 1991.
6. S. Lee, H. Ahn, "Fuzzy cognitive map based on structural equation modeling for the design of controls in business-to-consumer e-commerce web-based systems," Expert Systems with Applications, vol. 36, issue 7, pp. 10447-10460, Sep. 2009.
7. E. Papageorgiou, C. Stylios, P. Groumpos , "Fuzzy cognitive map learning based on nonlinear hebbian rule," AI 2003:
Advances in Artificial Intelligence Lecture Notes in Computer Science, vol. 2903, pp 256-268 , 2003.
8. B. Kosko, "Fuzzy cognitive maps," Int. J. Man-Machine Studies, vol. 24, pp. 65-75, (1986.
9. P. P. Groumpos, "Fuzzy cognitive maps: basic theories and their application to complex systems," in M. Glykas, Fuzzy Cognitive Maps: Advances in Theory, Methodologies, Tools and Applications, vol. 247, pp. 1-22, Springer-Verlag Berlin Heidelberg, 2010.
10. E. I. Papageorgiou, "Review study on fuzzy cognitive maps and their applications during the last decade," in Proc. 2011 IEEE Int. Conf. on Fuzzy Systems, IEEE Computer Society, Taipei, Taiwan, pp. 828-835, 2011.
11. P. Chytas, M. Glykas, G. Valiris, "A proactive balanced scorecard," Int. J. Information Management, vol. 31, issue 5, pp. 460-468, Oct. 2011.
12. D. Kardaras, G. Mentzas, "Using fuzzy cognitive maps to model and analyse business performance assessment," Advances in Industrial Engineering Applications and Practice II, pp. 63-68, 1997.
13. M. Glykas, "Fuzzy cognitive strategic maps in business process performance measurement," Expert Systems with Applications, vol. 40, issue 1, pp. 1-14, Jan. 2013.
14. G. Xirogiannis, M. Glykas, "Fuzzy causal maps in business modelling and performance-driven process reengineering," Methods and Applications of Artificial Intelligence Lecture Notes in Computer Science, vol. 3025, pp. 331341, 2004.
15. K. C. Lee, H. Lee, N. Lee, J. Lim, "An agent-based fuzzy cognitive map approach to the strategic marketing planning for industrial firms," Industrial Marketing Management, vol. 42, issue 4, pp. 552-563, May 2013.
16. C. Lopez, J. L. Salmeron, "Dynamic risks modelling in ERP maintenance projects with FCM," Information Sciences, vol. 256, pp. 25-45, Jan. 2014.
17. D. Kardaras, B. Karakostas, "The use of fuzzy cognitive maps to simulate the information systems strategic planning process," Information and Software Technology, vol. 41, issue 4, pp. 197-210, March 1999.
18. J. L. Salmeron, "Augmented fuzzy cognitive maps for modelling LMS critical success factors," Knowledge-Based Systems, vol. 22, issue 4, pp. 275-278, May 2009.
19. Z. Irani, A. Sharif, M. Mustafa Kamal, P. E.D. Love, "Visualising a knowledge mapping of information systems investment evaluation," Expert Systems with Applications, vol. 41, issue 1, pp. 105-125, Jan. 2014.
20. D. E. Koulouriotis, "Investment analysis & decision making in markets using adaptive fuzzy causal relationships," Operational Research, vol. 4, issue 2, pp. 213-233, May-Aug. 2004.
21. G. Xirogiannis, P. Chytas, M. Glykas, G. Valiris, "Intelligent impact assessment of HRM to the shareholder value," Expert Systems with Applications, vol. 35, issue 4, pp. 20172031, Nov. 2008.
22. Seyyed Yousef Oleyaei Motlagh, "The role of human resource management in increasing performance and implementation six sigma projects with using fuzzy cognitive maps," in Proc. 1th Int. Conf. Business Administration and Corporate Social Responsibility, Baku, pp. 1-15, Nov. 2012.
23. P. Macedo, L. M. Camarinha-Matos, "A qualitative approach to assess the alignment of Value Systems in collaborative enterprises networks," Computers & Industrial Engineering, vol. 64, issue 1, pp. 412-424, Jan. 2013.
24. I. Kang, S. Lee, J. Choi, "Using fuzzy cognitive map for the relationship management in airline service," Expert Systems with Applications, vol. 26, issue 4, pp. 545-555, May 2004.
25. C. Miao, Q. Yang, H. Fang, A. Goh, "A cognitive approach for agent-based personalized recommendation,"
Knowledge-Based Systems, vol. 20, issue 4, pp. 397-405, May 2007.
26. G. Xirogiannis, M. Glykas, "Intelligent modeling of ebusiness maturity," Expert Systems with Applications, vol. 32, issue 2, pp. 687-702, Feb. 2007.
27. S. M. Reza Nasserzadeh, M. Hamed Jafarzadeh, T. Mansouri, and B. Sohrabi, "Customer satisfaction fuzzy cognitive map in banking industry," Communications of the IBIMA, vol. 2, no. 21, pp. 151-162, 2008.
28. Z. Irani, A. M. Sharif, P. E.D. Love, "Mapping knowledge management and organizational learning in support of organizational memory," Int. J. Production Economics, vol. 122, issue 1, pp. 200-215, Nov. 2009.
29. C. H. Wang, S. C. Chen and K. Y. Chen, "Using fuzzy cognitive map and structural equation model for market-oriented hotel and performance," African J. Business Management, vol. 5 (28), pp. 11358-11374, Nov. 2011.
30. G. Buyukozkan, Z. Vardaloglu, "Analyzing of CPFR success factors using fuzzy cognitive maps in retail industry," Expert Systems with Applications, vol. 39, issue 12, pp. 1043810455, Sep. 2012.
ABOUT AUTHORS
NTARLAS, Orestis D, Graduate Student in the Dept. of Electrical and Computer Engneering, University of Patras, Patras, Greece.
GROUMPOS, Peter P., Professor in the Dept. of Electrical and Computer Engneering, University of Patras, Patras, Greece. Ph.D. (State University of New York at Buffalo, U. S. A., 1978) in Electrical and Engineering.
МЕТАДАННЫЕ
Название: Обзор приложений нечетких когнитивных карт в
бизнесе и управлении. Авторы: О. Д. Нтарлас, П. П. Грумпос Организация: Университет г. Патры, Греция. Email: ntarlas@ece.upatras.gr, groumpos@ece.upatras.gr Язык: английский.
Источник: Вестник УГАТУ. 2014. Т. 18, № 5 (66). С. 3-7,
ISSN 2225-2789 (Online), ISSN 1992-6502 (Print). Аннотация: Целью данной статьи является представить обзор существующих на сегодня приложений нечетких когнитивных карт управления бизнесом, определяющие свои сравнительные преимущества в традиционных экспертных системах. Таким образом, был составлен обзор исследований НКК, связанных с бизнесом и управлением, показана их эволюция на протяжении последнего десятилетия. Различные представительные приложения нечетких когнитивных карт проиллюстрированы в различных областях управления бизнесом, таких как электронный бизнес, оценка эффективности, управление людскими ресурсами, планирование и инвестиционные процессы принятия решений. Ключевые слова: нечеткие когнитивные карты; электронный бизнес; оценка эффективности; управление.
Об авторах:
НТАРЛАС Орестис Д., асп. ф-та э/техн. и выч. энергетики.
ГРУМПОС Петер П. проф. ф-та э/техн. и выч. энергетики. PhD в обл. э/техн. и техн. (Гос. ун-т Нью-Йорка в Буффало, США, 1978).
