성과지형과 탐색 방식 간의 부조화

Abstract

Organizational scholars proposed performance landscape models to conceptualize the relationship between organizational choices and performance. Since its inception, scholars have introduced various performance landscapes and have examined performance implications by adopting the appropriate search methods for each performance landscape. However, the assumption that the searching agent is aware of and able to select the most appropriate search method is beyond the level of rationality usually adopted in the management field. In this study, we introduce and review various types of performance landscapes studied in the field of management (smooth performance landscape, single-peak performance landscape, and multi-peak performance landscape) and the search methods mainly adopted (local search, long-distance search, and Q-learning algorithm-based search) for each landscape. Taking the assumption that the searching agent does not know the appropriate search method for a given landscape, we conduct a thought experiment on misfits between performance landscapes and search methods. In many cases, when there is a misfit between the performance landscape and the search method, a decrease in performance is expected. However, performance increases are also possible. Based on these expected patterns, possible future research directions are proposed.