Modelling Organizational Learning in Complex and Turbulent Environments: an Agent-Based Simulation in NK Landscapes

Organizations are increasingly considered Complex Adaptive Systems (CAS). This view has outlined the need to adopt a different theoretical and methodological approach in organizational research and design. The present paper copes with this challenge and aims firstly at introducing an agent-based computational laboratory, named CLOD (Computational Laboratory of Organizational Design), to explore the advantages that agent-based approaches could offer to scholars and practitioners in the organization research, particularly in the field of organizational design. Furthermore, some generative simulation experiments are reported and discussed to analyze the impact of the increasing complexity of the external environment on learning performances of organizations of different sizes. The conceptual architecture of the computational laboratory CLOD is based on the seminal March's model of exploration and exploitation in organizational learning. Some generative experiments have been performed (18 experimental sets), in which organizations of different sizes have to learn by coping with external environments characterized by increasing complexity levels (modeled through the values of the parameter K in the NK landscape), and, eventually, by a certain level of turbulence. In a stable setting, as expected, the learning performances of the organization decrease significantly, moving from a single peak environment (K=0) to higher levels of environmental complexity (K =5). Furthermore, the increasing size of the organization is related to better performances in stable environments for low or moderate levels of complexity. In turbulent and complex environments, instead, the size of the organization, particularly in the medium-long period, seems to be not influent and other organizational solutions have to be found to reach significant learning performances. The Clod lab can be considered, at this stage of the research, as a tool supporting theory development or refinement in organizational design disciplines. Further applications can be tested for practical and teaching purposes.