Double-loop learning in project environments: An implementation approach

Project-based learning is based on the idea of iteratively learning for future projects from the successes and failures of past projects. This paper proposes a semi-automated implementation approach for double-loop learning in project environments. A combined application of two complementary methods is suggested for this purpose: Latent Semantic Analysis (LSA) and Analytic Network Process (ANP). By this means, the approach addresses two problems of the project management practice. First, the information overload in project environments, whereby the LSA is used for the semi-automated extraction of lessons learned from large collections of textual project documentation. Second, the lack of procedures and methods for the practical implementation of available project knowledge, whereby the ANP is used for the systematic modeling of extracted lessons learned and their integration into the evaluation of project concepts and current project management routines. Thus, the proposed implementation approach improves the ability of project-based organizations to consequently learn from past failures or successes. From a practical perspective, evident shortcomings of existing computerized double-loop learning approaches are addressed. The proposed approach contributes to the project management practice not only by demonstrating a solution for the exploration of representative and potentially new lessons from multiple combined experience reports, but also by presenting a solution for the systematic assessment of such project-governing variables and their mutual relationships as part of the decision-making in new projects. From a theoretical perspective, specific research avenues for further development of the double-loop learning concept by means of expert and intelligent systems are provided.