A game theoretic analysis of knowledge sharing behavior of academics: Bi-level programming application. (May 2019)

• Record Type: Journal Article
• Authors: Tabatabaei, Mohsen; Afrazeh, Abbas; Seifi, Abbas
• Is Part Of: Computers & industrial engineering. Volume 131(2019)
• Page Start: 13
• Page End: 27
• Publication Date: 2019-05
• Subjects: Bi-level programming -- Game theory -- Knowledge sharing behavior -- Academics -- Higher education -- Motivation-opportunity-ability framework; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2019.03.031
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: A bi-level model is provided to examine knowledge sharing behavior of academics. The bi-level model has a feasible solution based on two presented propositions. Numerical results indicate that increasing compensation is not always a best solution. Effect of trust level on the performance depends on priorities of faculty members. Abstract: Knowledge sharing amongst academics is a key process for universities to survive in the face of new changes in the education market. Despite the bulk of research on knowledge sharing, there are few studies on knowledge sharing analysis in higher education based on the game theory. Considering this gap, this paper aims to propose a nonlinear Bi-Level Programming (BLP) model to analyze the knowledge sharing behavior of academics. The proposed model considers decisions made by Faculty Head (FH) and Faculty Members (FMs) at two separate but integrated levels. The upper-level problem is related to FH decisions about compensation rates, and the lower-level problem represents FM decisions in the motivation, opportunity and ability (MOA) framework. This modeling approach has several advantages including the consideration of both FHs and FMs decisions, and various factors influencing knowledge sharing using the MOA framework in a single model as well as the analysis of four different activities of FMs, including teaching, learning, sharing tacit knowledge and publishing codified knowledge in a single integrated model. The bi-level programming model was reduced to a single level nonlinear problem using KKT conditions. The resultant problem was solved for a set of randomly generated data. Results indicated that the optimal behavior of an FM varied relative to characteristics of other FMs. Also, the analysis of the solution at different levels of trust showed that its improvement might have a different effect on FMs. Extending the application of the model and developing efficient algorithms for its solution are suggested for future research.