Enhancing the Strategic Decision-Making Capabilities of Construction Project Managers Through Reinforcement Learning: A Simulation-Based Framework Under Economic Uncertainty

Enhancing the Strategic Decision-Making Capabilities of Construction Project Managers Through Reinforcement Learning: A Simulation-Based Framework Under Economic Uncertainty

Mahmoud Golabchi1 Mohammad Hossein Mohseni2

1) Professor, College of Fine Arts, University of Tehran
2) Pars University of Art and

IOI Code : ABGG-ACAGGZ
Publication : 11th.International Conference on Researches in Science & Engineering & 8th.International Congress on Civil, Architecture and Urbanism in Asia(11icrsie.ir)
Abstract :
    Construction projects face numerous strategic decision-making challenges due to their substantial capital requirements, long execution periods, and vulnerability to economic fluctuations. Under such conditions, conventional project planning and control approaches alone are insufficient to address the need for continuous strategic reassessment and the evaluation of the long-term consequences of decisions. This study aims to investigate the potential of reinforcement learning to enhance the strategic decision-making capabilities of construction project managers and to propose an integrated simulation-based framework under economic uncertainty. Adopting a conceptual approach grounded in the analysis of theoretical foundations, the study develops a seven-layer framework comprising strategic governance, data monitoring, environment simulation, reinforcement learning problem formulation, agent training and evaluation, human–system interaction, and organizational learning. Within this framework, the learning agent develops decision-making policies across various economic and operational scenarios through interaction with a simulated environment and evaluation of the consequences of its actions. This approach can be applied to cash flow management, resource allocation, activity prioritization, and the evaluation of alternative strategies. Nevertheless, its effectiveness depends on simulation validity, data quality, reward function design, and human oversight. The proposed framework provides a conceptual foundation for developing intelligent decision support systems; however, empirical validation is necessary to determine its practical effectiveness.

Keywords : reinforcement learning strategic decision making construction project management economic uncertainty simulation artificial intelligence decision support