UTM. Evaluation of Knowledge Management Processes using Fuzzy Logic. Kuan Yew Wong, PhD UNIVERSITI TEKNOLOGI MALAYSIA

Transcription

1 UTM UNIVERSITI TEKNOLOGI MALAYSIA Evaluation of Knowledge Management Processes using Fuzzy Logic Kuan Yew Wong, PhD

2 UTM UNIVERSITI TEKNOLOGI MALAYSIA INTRODUCTION

3 Knowledge - One of the most important assets of any organization Managing knowledge has been part of our daily working routines or activities KM processes - Activities that have or involve the elements of KM

4 Knowledge Acquisition Knowledge Creation and Generation KM Processes Knowledge Application and Utilization Knowledge Transferring and Sharing Knowledge Codification and Storing

5 UTM UNIVERSITI TEKNOLOGI MALAYSIA PROBLEM STATEMENT

6 To keep track and improve a company s KM processes They have to be evaluated and improved continuously Daunting task due to their subjective and intangible nature

7 Vagueness of indicators or metrics Fuzziness of data Insufficient and incomplete data

8 UTM UNIVERSITI TEKNOLOGI MALAYSIA OBJECTIVE

9 To propose an evaluation approach using fuzzy logic Able to cope with vagueness and uncertainties Able to deal with incomplete and fuzzy data As an alternative evaluation method for companies

10 UTM UNIVERSITI TEKNOLOGI MALAYSIA LITERATURE REVIEW

11 Fuzzy Logic Created by Lofti Zadeh in 1965 Applied in many fields such as control theory, image processing, AI, management, etc. Deals with reasoning that is approximate rather than fixed and exact Able to handle the concept of partial truth

12 Easy to understand Why Fuzzy Logic? Tolerant of imprecise data Does not require precise numerical information input Fuzzy if-then rules resemble human thinking

13 UTM UNIVERSITI TEKNOLOGI MALAYSIA METHODOLOGY

14 Knowledge Acquisition Knowledge Creation and Generation Evaluation Scope Knowledge Transferring and Sharing Knowledge Codification and Storing Knowledge Application and Utilization

15 Development of Evaluation Approach MATLAB - Fuzzy Logic Toolbox Provides extensive debugging and optimization features Graphical user interface (GUI) environment

16 Building a Fuzzy Logic Approach in Five Steps 1. Fuzzification of inputs Convert inputs into membership values or degrees of membership through a membership function The triangular membership function is used

17 2. Application of fuzzy operators Resolve the antecedent to a single number between 0 and 1, which is the degree of support for the rule Three types of fuzzy operators: AND (minimum) OR (maximum) NOT (fuzzy complement)

18 3. Application of the implication method The input for the implication process is a single number given by the antecedent The output is a fuzzy set Implication is implemented for each rule In Mamdani s max min mechanism, implication is modeled by the minimum(and) operator

19 4. Aggregation of all outputs All the rules must be combined in order to make a decision Aggregation process combines all the fuzzy sets that represent the outputs of each rule into a single fuzzy set In Mamdani s max min mechanism, the outputs of each rule are combined using the maximum (OR) operator

20 5. Defuzzification Defuzzify the aggregated output value into a single crisp value (performance value of KM processes) The centroid defuzzification method is used

21 UTM UNIVERSITI TEKNOLOGI MALAYSIA EXAMPLE

22 Constructs Knowledge acquisition Knowledge creation Knowledge utilization Knowledge codification and storing Knowledge transferring and sharing Example of Metrics for Evaluation Metrics a 1. The frequency of employees attending training or seminars to acquire knowledge. a 2. The frequency of employees accessing the company s knowledge repositories to acquire knowledge. b 1. The frequency of employees participating in brainstorming sessions to create new knowledge. b 2. Number of new knowledge, ideas and solutions created. c 1. The frequency of employees applying useful proposals or ideas in practice. c 2. The frequency of employees applying knowledge to solve problems. d 1. Amount of time spent codifying and storing knowledge into the company s knowledge repositories. d 2. Employees level of willingness to contribute to the company s knowledge repositories. e 1. The frequency of employees involving in work related discussion to share knowledge. e 2. The frequency of employees having meetings.

23 Linguistic Terms and Their Input Scores Linguistic Terms Scores None 0 Extremely Low 10 Very Low 20 Low 30 Slightly Low 40 Average 50 Slightly High 60 High 70 Very High 80 Extremely High 90 Perfect 100

24 Example Constructs Input Scores for Metrics Knowledge acquisition (50, 40) Knowledge creation (40, 30) Knowledge utilization (60, 20) Knowledge codification and storing Knowledge transferring and sharing (40, 70) (30, 40)

25 Fuzzy Set of a1, a2, b1, b2, c1, c2, d1, d2, e1 and e2 Linguistic Variables Interval Very low (VL) (0,0,25) Low (L) (0,25,50) Average (A) (25,50,75) High (H) (50,75,100) Very high (VH) (75,100,100)

26 Example of Inference Rules of a1 and a2 a1 a2 VL L A H VH VL VB VB B B A L VB B B A G A B B A G G H B A G G VG VH A G G VG VG

27 Fuzzy Set of Output Linguistic Variables Interval Very bad (VB) (0,0,25) Bad (B) (0,25,50) Average (A) (25,50,75) Good (G) (50,75,100) Very good (VG) (75,100,100)

28 Active Rules and Performance Value of Knowledge Acquisition

29 Evaluation Results from the Example Constructs Input scores Output scores Knowledge acquisition (50, 40) 39.5 Knowledge creation (40, 30) 31.7 Knowledge utilization (60, 20) 35.5 Knowledge codification and storing Knowledge transferring and sharing Total average score (40, 70) 57.6 (30, 40)

30 UTM UNIVERSITI TEKNOLOGI MALAYSIA DISCUSSION

31 Formulate improvement strategies based on the evaluation results Knowledge creation and knowledge transferring and sharing have the lowest score To improve knowledge creation conduct more brainstorming sessions to create new ideas, products or solutions

32 To improve knowledge transferring and sharing improve organizational culture to promote collaboration and interaction among employees In practice, avoid using a large number of metrics for each construct Fuzzy if-then rules will increase exponentially based on the number of inputs

33 The membership functions, fuzzy sets, fuzzy rules, and the number of metrics used can be changed and fine tuned In advanced cases, a weight system can be added

34 UTM UNIVERSITI TEKNOLOGI MALAYSIA CONCLUSION

35 Fuzzy logic as an evaluation method Involves five simple steps: 1. Fuzzification of inputs 2. Application of fuzzy operators 3. Application of the implication method 4. Aggregation of all outputs 5. Defuzzification

36 Effective evaluation of KM processes helps to: Keep track of KM related activities Provide constructive information for decision making and improvement Ensure KM initiatives are successful Obtain better organizational performance and competitive advantage

37 UTM UNIVERSITI TEKNOLOGI MALAYSIA