KNOWLEDGE FLOW NETWORKS

KNOWLEDGE FLOW NETWORKS AND COMMUNITIES OF PRACTICE FOR KNOWLEDGE MANAGEMENT
RAJIV KHOSLA & MEI-TAI CHU
Research Centre for Computers, Communication and Social Innovation School of Management, La Trobe University, Melbourne, Vic 3086, Australia
E-mail: r.khosla@latrobe.eud.au
K. G. YAMADA, S. DOI, K. KUNEIDA & S. OGA
C&C Innovation Res. Labs, NEC Corporation
8916-47 Takayama-cho, Ikoma-Shi, Nara, 630-0101, Japan.
E-mail: kg-yamada@cp.jp.nec.com
This research discusses KFNs in the context of Communities of Practice (CoPs) and Knowledge Management (KM). KFNs unlike workflow can often transcend organizational boundaries and are distinct and different than workflow models. CoPs involve both personal and organizational aspects, and are an iteration of the transmission between explicit and tacit knowledge. This research develops, implements, and analyzes a CoPs Centered KFNs model in a multinational organization. The CoPs Centered KFNs model is underpinned in a CoPs model built around four organization performance evaluation dimensions and sixteen criteria. Many criteria and comprehensive segments should be taken into consideration while establishing CoPs model, this explains why this research employs fuzzy multi-criteria decision making. The cluster analysis techniques are used for evaluation of the CoPs Centered KFN model. The result of attribute analysis via KFNs model has been designed to determine the characteristic of each cluster and identify suggestions for effective linkage among knowledge workers.
1. Introduction
KFNs not only fall within the scope of managers, information technologists and knowledge workers but involve CoPs in organization learning (Lesser, 2001). Most of the existing work on knowledge flow networks has centered around linking people based on organization structure, tasks, and knowledge compatibility (Zhuge, 2006). This research proposes to enhance in design of KFN by modeling them based on CoPs in organization learning. In CoPs, like in KFNs, people with a common goal come together to create, learn, process and share knowledge and learning based on best practices. A CoPs model has been defined in this study, which constitutes 16 criteria along four performance measurement dimensions. These criteria and dimensions are used to identify common interaction factors (beliefs and attitudes) which link and facilitate effective knowledge sharing and learning between knowledge workers in KFNs. These factors and the CoPs model have been validated using a large multinational organization as a case study.
The research is organized as follows. Section 2 covers the theoretical considerations underpinning the definition and construction of KFNs model. Section 3 describes implementation and Techniques of KFNs model based on survey of R&D personnel in a multinational organization to enhance organization learning. Section 4 concludes the research.
2. KFNs Model
In this section KFNs model based on CoPs is constructed. The assumptions are made that design of KFNs is driven by the need to develop effective knowledge sharing and knowledge management (KM) mechanisms in order to enable organizations to compete in a knowledge based economy. In this context, firstly, defining a CoPs model, the criteria and dimensions it is based and the business strategies or benefits which can be evaluated using the model. The work is followed with definition of CoPs centered KFNs model which is used for implementation and analysis in this case study.
2.1. CoPs Context and its Benefits
Despite the rise of technology-based Knowledge Management tools, implementations often fail to realize their stated objectives (Ambrosio 2000). It is envisaged that 70% of existing knowledge management tools have failed to achieve the anticipated business performance outcomes they had been designed for (Malhotra 2004). One of the primary reasons identified for the failure of existing KM tools has been that existing Knowledge Management tools and research have primarily been designed around technology-push models as against strategy-pull models (Malhotra 2004). The technology-push model which is based on application of information technologies on historical data largely produce pre-specified meanings/knowledge and pre-specified outcomes which are useful in predictable and stable business environments. On the other hand, strategy-pull model turns the technology-push model on its head and drive the construction and creation of knowledge and related actions based on business strategy and performance driven outcome rather than somehow find a business strategy fit for the pre-specified knowledge and outcomes produced by technology-push models.
In an era where organizations are undergoing rapid, discontinuous and turbulent change it is imperative that KM systems and organizational entities like CoPs which facilitate KM and organizational transformation are more closely aligned with business strategies and goals of an organization. This would enable organizations to respond more quickly to changing business environments and business process and corresponding change in their KM needs from time to time.
Wenger (1998) first proposed CoPs in the Harvard Business Review, who believes CoPs is an informal group sharing knowledge, points out CoPs is composed by three critical elements (mutual engagement, joint enterprise, shared repository). Allee (2000) thought knowledge should include and utilize CoPs to create organizational knowledge. Besides, CoPs are distinguishing from other organizational groups such as formal divisions, project teams and informal network (Cohendet & Meyer-Krahmer, 2001; Allee, 2000; Wenger et al., 2002). CoPs can enable member interaction, knowledge sharing, organization learning, and open innovation simultaneously; it emphasizes more on facilitating, extracting and sharing tacit knowledge to maximize KM value. Many world class companies have taken CoPs as a new central role in the value chain (Chu et al, 2007).
Acknowledging that CoPs can link with organizational performance very well, CoPs are essential to overcome the inherent problems of a slow-moving traditional hierarchy in a fast-moving knowledge economy. Therefore, this research uses the four CoPs benefits or business strategies Induce Innovation Learning, Promote Responsiveness, Increase Core Competency, and Enhance Working Efficiency to develop the CoPs model as shown in Figure 1. These four CoPs business strategies need to be well defined and then pursued, because they will influence the KM achievements and the community’s resources allocation direction.




















Figure 1: CoPs Centered Evaluation Hierarchy Model

The first benefit is to Induce Innovation Learning. The specific characteristics include cross-domain sharing to support new idea and creation according to common interests through group learning. The CoPs under this strategy often provide a safe or low-cost infrastructure for try and error attempts freely to facilitate new thinking and innovation.
The second benefit is to Promote Responsiveness by collecting and classifying knowledge objects. CoPs can directly obtain the problem-oriented solution, because the colleagues with similar working experiences are easy to find. They can help other members who are facing same questions based on the common language and shared foundations which lead to promote responsiveness.
The third benefit is to Increase Core Competency. Members can promote skill by shifting the best knowledge practices. It will be efficient to figure out who are domain-experts, how to enable insight exchange between senior and junior members. The organization principals can be established and increase core competency.
The fourth benefit is to Enhance Working Efficiency. CoPs members can reuse existing intellectual property invented by others in a well structured database easily, access related documents and authors’ information quickly. The entire productivity will be improved and working efficiency will be enhanced in a disciplined way.
2.2. CoPs Model and its Components
In order to realize the four business benefits or strategies the CoPs model is defined and evaluated along four performance dimensions and sixteen criteria as outlined in Chu et al, (2007). The four dimensions are explained as follows respectively:
- Locus of Leadership: relates to enforcement or volunteer, wholly or partially adoption
- Incentive Mechanism: relates to award or punishment
- Member Interaction: relates to sharing or security
- Complementary Asset: relates to infrastructure and resource
The Locus of Leadership dimension contains four criteria: Top-Down Assigning, Bottom-Up Teaming, Total Execution, and Partial Pilot run. The Incentive Mechanism dimension contains: Substantive Reward, Psychological Encourage, Achievements Appraisal Basis, and Peers Reputation. The Member Interaction contains: Homogeneity of members, Differential members, Emphasize security, and Emphasize cross-domain Sharing. The Complementary Asset dimension contains: Give Extra Resources, Just Daily Work, Integrated IT Platform, and Independent IT platform.
2.3. CoPs Centered Knowledge Flow Network Model
In the preceding section the ground related to definition and construction of CoPs model has been outlined. In this section CoPs cantered parameters are used to define the components and terminologies of the knowledge flow network model. The KFN includes quantitative implications of the human and social factors like beliefs and attitudes for interaction between knowledge workers derived from the CoPs model (Thomas et al., 2001). These interaction beliefs and attitudes for knowledge sharing are based on the sixteen criteria used by the CoPs model. KFN can also be considered as CoPs in an organization where people with a common goal come together to create, learn, process and share knowledge based on best practices. Organizations and research teams are held together by CoPs or KFNs.
The purpose for designing a KFN model in this research is to develop actual human networks which can then be used for creation, learning, processing and sharing of knowledge (Davenport et al., 2004; Malhotra, 2004; Ratcliffe et al., 2000; Nissen, 2002; Nonaka, 1994; Thomas et al., 2001; Zhuge, 2003; Desouza, 2003). Knowledge especially that resulting from innovation needs, is regarded as an organizational transformation issue. It involves transmission of explicit, tacit and embodied knowledge in an iterative manner through KFN.





















Figure 2: Knowledge Flow Network Model
A KFNs model as shown in Figure 2 consists of knowledge nodes (human or knowledge portal or process), knowledge links and weight which help to specify the strength of the knowledge link. With the definition of CoPs and existing research, knowledge workers share knowledge based knowledge compatibility as well as a set of interaction principles and beliefs which define their underlying knowledge sharing philosophy (Thomas et al., 2001). Although these interaction principles are not a determining factor for knowledge sharing they do influence the effectiveness and efficiency of knowledge sharing between two knowledge workers. These interaction principles and beliefs are defined based on the 16 criteria defined in CoPs model.
To draw an analogy, consider a situation for recruitment of sales person for selling computers. On one hand the recruitment panel will determine the knowledge compatibility of the sales person in the domain of computers. On the other hand, they will also study or analyze (based on range of criteria) how this sales person will interact with a customer in an actual selling situation. Similarly, knowledge level and space of a knowledge worker or a researcher can be determined based on their experience, CV, etc. However, to what extent they actually engage in knowledge sharing (especially, tacit knowledge) may be influenced by the 16 criteria for knowledge sharing and management. Other factors which can influence knowledge sharing can be trust and psychological profiles of the cooperating knowledge workers. However, the latter factors can be extended based on this research.
Therefore in Figure 2 we consider two types of weights, knowledge space weight and interaction principles weight. The knowledge space weight can vary between 0-1 and can be specified by the group or network leader based on knowledge and experience of the two knowledge workers, between discussions and consensus to calculate the impact of interaction principles on the overall effectiveness and efficiency of the knowledge link between two human nodes.
Thus knowledge link weight between two human nodes is calculated as follows:
KLWmn = KSWmn + ∑i=16i=1 CCWmni
Where KLWmn is the knowledge link weight between nodes m and n, KSWmn is the knowledge space weight between nodes m and n, and CCWmni is the common criteria weight of criteria i between nodes m and n. The criteria weight is normalized between 0-1. The criteria with weight 0.1 or above may be added to determine CCWmn. However, the knowledge flow pattern is different than work flow and may or may not follow the same pattern or path as the work flow.
KFNs model can also assist in formation and growth of knowledge flow teams for R&D organizations as well as identification of high and low knowledge energy nodes. The human node with the highest number of links is the node with highest knowledge energy as it represents knowledge sharing and interaction potential of the node.
3. Techniques and Implementation of CoPs Centered KFN Model
In this section the authors describe the techniques used for construction of CoPs Centered knowledge flow network model in a large multinational organization. These techniques include a CoPs questionnaire based survey of knowledge workers. The survey is used to evaluate the importance attributed by knowledge workers to 16 CoPs criteria of knowledge workers along four business performance evaluation dimensions. Fuzzy MCDM (Multi-Criteria Decision-Making) techniques are used to calculate the importance attributed to each dimension and each criterion by the knowledge worker participating in the survey. Finally clustering technique is used to connect knowledge workers with common criteria (attitudes and beliefs) in CoPs centered KFNs model. Intuitively, common attitudes and beliefs between two knowledge workers imply that knowledge sharing among them is likely to be more effective than between knowledge workers with dissimilar attitudes and beliefs. The common criteria between two knowledge workers in KFNs are also used to determine strength of CCWmn link between knowledge workers in a KFN. The implementation was conducted in one large R&D organization, and the questionnaire was distributed to a broad sampling of researchers, to seek their views and calculate their final values. The aim is to provide a valuable reference when choosing suitable CoPs business strategies. Thirty nine valid questionnaires out of seventy were collected with a response rate at 55.7%.
The attribute analyses of KFN Model for CoPs designs can determine the characteristic of each cluster and identify suggestions for effective linkage. This KFN model adopted cluster analysis to be the basis of attribute analysis. Based on the differences of each participant, a hierarchical cluster diagram is generated. The similarity degree increased gradually from top down; the lower the knowledge workers are on the hierarchy, the more unique they appear to be (Pellitteri, 2002; Akamatsu et al., 1998; OECD, 1996).
The cluster analysis contains several steps. First, we input the factor scores to the model of cluster analysis. Second, we divided five clusters among all the participants. Third, we calculated the mean value and variable number of score of factor for each knowledge worker so as to explain their differences and characteristics. This research divided into five groups after the analysis results and actual discussions about the features towards CoPs beliefs. Table 1 demonstrates the participant distribution of five clusters.





Table 1: Participants in each Knowledge Flow Network
Knowledge Flow Network Number No. of People
1 9
2 7
3 5
4 9
5 9
4. Results
In order to illustrate the application of knowledge link weight, the KFNs of the case study is shown in Figure 3. The knowledge flow network has been constructed using the CoPs Centred model designed in previous section. The CoPs model is used to design a questionnaire involving four dimensions, sixteen criteria, and four business strategies or performance preferences. The sixteen criteria represent among other aspects, represent beliefs and interaction principles of knowledge workers for knowledge sharing and management.
The feedback from 39 participants is used to compute the weight or relative importance assigned to each criterion by a participant. The weight values were than used to cluster the weighted responses from 39 participants. The purpose of clustering is to determine the similarities in relative importance of sixteen criteria among 39 participants. The clustering technique was derived from SPSS software (Zadeh, 1981). Five clusters or groups of researchers are identified. Each group or cluster in this research is considered to be eligible to form a knowledge flow network.













Figure 3: Sample Knowledge Flow Network for Number 5.
Table 2 shows similar weight values for various criteria allocated by members of 5 KFNs. The weight values above 0.1 are highlighted in bold. These are used to calculate the Common Criteria Weight (CCW) between two members in KFNs. The criteria weights for criteria differential member and cross-domain sharing are added up. The values based on experience of members/researchers in a related knowledge domain have been used for illustration purpose only.


Table 2: Common Criteria Weight and Knowledge Flow Network Number

5. Conclusion
Organizations in this research are viewed as KFNs involved in knowledge creation, knowledge sharing and innovation. This is in contrast to the traditional view that organizations consist primarily of workflow networks. KFNs consist of knowledge nodes, knowledge links and knowledge link weight respectively. The knowledge nodes are primarily human nodes but also can be resource nodes (e.g., robot, knowledge portals, databases, WWW).
The significant goal of this research is to study how CoPs can help to synergize the existing collaboration and construction of KFN in an open innovation infrastructure. The KFN are constructed based on actual study of CoPs in one large multinational R&D organization. The knowledge link weight between two human nodes consists of Knowledge Space (or compatibility) Weight (KSW) and Common Criteria Weight (CCW). KSW between two human nodes is determined by a manger or group leader based on CV, experience of the two human nodes and their knowledge compatibility. The 16 criteria along four Performance evaluation dimensions (Locus of leadership, Member interaction, Incentive mechanism and Complementary asset) in the CoPs questionnaire, among other aspects, can be considered to provide information on the interaction attitude and beliefs of researchers for cooperation and knowledge sharing. These interaction principles although, not a determining factor for knowledge sharing, can improve or enhance the effectiveness of knowledge sharing, creation and innovation. The feedback on the CoPs questionnaire, among other aspects, is used for clustering the researchers in knowledge flow network group. In this research 39 participants have been clustered into five KFNs. Organization can implement CoPs as a major approach to outline the future roadmap by frequent member interaction. Thus findings of this research can promote performance and can facilitate allocation of organizational resources for knowledge sharing and innovation among the participants.
References
Allee, Verna. (2000), Knowledge Networks and Communities of Practice, OD Practitioner Online, 32(4)
Ambrosio, J. (2000), Knowledge Management Mistakes, http://www.computerworld.com/industrytopics/energy/story/0,10801,46693,00.html
Buckley, J. J. (1985), Ranking Alternatives Using Fuzzy Number, Fuzzy Sets and Systems, 15(1), 21-31.
Chu M. T., Shyu J. Z., Tzeng G. H. & Khosla R. (2007), Comparison among Three Analytical Methods for Knowledge Communities Group-Decision Analysis, Expert Systems with Applications, 33(4), 1011-1024.
Chu M. T., Shyu J. Z., Tzeng G. H. & Khosla R. (2007), Using Non-Additive Fuzzy Integral to Assess Performance of Organization Transformation via Communities of Practice, IEEE Transactions on Engineering Management, 54(2), 1-13.
Clatworthy, J., Buick, D., Hankins, M., Weinman, J., & Horne, R. (2005), The use and reporting of cluster analysis in health psychology: A review, British Journal of Health Psychology 10: 329-358.
Cohendet P. & Meyer-Krahmer F. (2001), The Theoretical and Policy Implications of Knowledge Codification, Research Policy, 30, 1563-1591.
Davenport, T. H., Jarvenpaa, S. I., Beer, M.C. (2004), Improving Knowledge Work Process, Sloan Management Review, 34 (4), 53-65.
Desouza, K.C. (2003), Facilitating Tacit Knowledge Exchange, CACM, 46(6), 85-86.
Fowlkes E. B. & Mallows C. L. (1983), A Method for Comparing Two Hierarchical Clusterings, Journal of the American Statistical Association, 78 (383): 553–584.
Genrich, A. & Lautenbach, K. (1981), System Modelling with High Level Petri Nets, Theoretical Computer Science, 35, 1-41.
Huang, Z. (1998), Extensions to the K-means Algorithm for Clustering Large Datasets with Categorical Values, Data Mining and Knowledge Discovery, 2, 283-304.
Hwang, C. L. & Yoon, K. (1981), Multiple Attribute Decision Making：Methods and Applications, A State-of-Art Survey, Springer-Verlag, New York.
Jensen, K. (1981), Colored Petri Nets and the Invariant Method, Theoretical Computer Science, 14, 317-336.
Kerzner, H. (1989), A System Approach to Planning Scheduling and Controlling, Project Management, New York, 759-764
Lesser E. L. & Storck J. (2001), Communities of Practice and Organizational Performance, IBM Systems Journal, 40(4)
MacKay, David J.C. (2003), Information Theory, Inference, and Learning Algorithms, Cambridge University Press
Malhotra, Y. (2004), Why Knowledge Management Systems Fail? Enablers and Constraints of Knowledge Management in Human Enterprises, American Society for Information Science and Technology Monograph Series, 87-112.
Michael J. Lyons, Shigeru Akamatsu, Miyuki Kamachi & Jiro, (1998), Coding facial expressions with Gabor wavelets Automatic Face and Gesture Recognition, Third IEEE International Conference, 200-205, 4-16 April.
Mon, D. L., Cheng, C. H. & Lin, J. C. (1994), Evaluating Weapon System Using Fuzzy Analytic Hierarchy Process Based on Entropy Weigh, Fuzzy Sets and Systems, 61, 1-8.
Nissen, M. E. (2002), An Extended Model of Knowledge Flow Dynamics, CACM, 8, 251-266.
Nonaka, I. (1994), A Dynamic Theory of Organizational Knowledge Creation, Organizational Science, 5(1), 14-37.
OECD 1996: The Knowledge Based Economy. Science, Technology and Industry Outlook, Paris.
OECD 1999: Measuring Knowledge in Learning Economies and Societies. Draft Report on Washington Forum.
Pellitteri, J. (2002), The Relationship between Emotional Intelligence and Ego Defence Mechanisms, The Journal of Psychology, 136, 182-194, March.
Petri, C. (1966), Communication in Automata, in Technical Report RADC-TR-65-377, 1, Rome Air Development Center, Griffths Air Base, USA, January.
Ratcliffe-Martin V., Coakes E., & Sugden G., (2000), Knowledge Management Issues in Universities, Vine Journal, Dec 121, 14-19, ISSN: 0305-5728.
Saaty, T. L. (1977), A Scaling Method for Priorities in Hierarchical Structures, Journal of Mathematical Psychology, 15(2), 234-281.
Saaty, T. L. (1980), The Analytic Hierarchy Process, New York, McGraw-Hill.
Tang, M. T. & Tzeng, G. H. (1999), A Hierarchy Fuzzy MCDM Method for Studying Electronic Marketing Strategies in the Information Service Industry, Journal of International Information Management, 8(1), 1-22.
Thomas, J.C., Kellog, W.A., & Erickson, T. (2001), The Knowledge Management Puzzle: Human and Social factors in Knowledge Management, IBM Systems Journal, 40 (4), 863-884.
Tsaur, S. H., Tzeng, G. H. & Wang, K. C. (1997), Evaluating Tourist Risks From Fuzzy Perspectives, Annals of Tourism Research, 24(4), 796-812.
Tzeng, G. H. & Shiau, T. A. (1987), Energy Conservation Strategies in Urban Transportation: Application of Multiple Criteria Decision-Making, Energy Systems and Policy, 11(1), 1-19.

Tzeng, G. H. & Teng, J. Y. (1994), Multicriteria Evaluation for Strategies of Improving and Controlling Air-Quality in the Super City: A case of Taipei city, Journal of Environmental Management, 40(3), 213-229.
Tzeng, G. H. (1977), A Study on the PATTERN Method for the Decision Process in the Public System, Japan Journal of Behavior Metrics, 4(2), 29-44.
Tzeng, G. H., Shian, T. A. & Lin, C. Y. (1992), Application of Multicriteria Decision Making to the Evaluation of New Energy-System Development in Taiwan, Energy, 17(10), 983-992.
U.S. Department of Commerce (1965), National Technical Information Service, NASA, PATTERN Relevance Guide, 3.
Wenger, E. (1998), Communities of Practice, Cambridge University Press.
Wenger, E., McDermott R. A., & Snyder W. (2002), Cultivating Communities of Practice, Boston: Harvard Business School Press.
Zadeh, L. A. (1981), A Definition Of Soft Computing, http://www.soft-computing.de/def.html.
Zhuge, H. (2003), Component-based Workflow Systems Design, Decision Support Systems, 35 (4), 517-536
Zhuge, Hai. (2006), Knowledge Flow Network Planning and Simulation, Decision Support Systems, 42, 571-592