MR-based dynamic optimization and simulation method for component-level construction of prefabricated buildings

To address the efficiency bottlenecks in multi-component collaborative construction of prefabricated buildings and the inadequacy of traditional manual scheduling in responding to dynamic on-site conditions, this study proposes an integrated optimization framework combining genetic algorithms (GAs) and mixed reality (MR). By developing a genetic algorithm-based optimization model, the automated component-level scheduling under multi-constraint conditions is achieved. The system incorporates a dynamic “perception–decision–optimization” loop to enhance schedule robustness against on-site disruptions. Innovatively integrating mixed reality technology, we establish a cyber-physical simulation system enabling full-scale component visualization through three-dimensional (3D) spatial mapping and real-time registration. Validated through a Beijing prefabricated building project, the framework demonstrates 27.28% efficiency improvement over conventional methods by achieving millimeter-level spatial alignment between digital plans and physical environments. Key innovations include: (1) a dual-driven optimization mechanism combining algorithmic scheduling with MR-enhanced verification, (2) an in-situ visualization of construction sequences through holographic component projection, and (3) a closed-loop process reconfiguration capability responsive to field variations. The empirical results confirm the system’s effectiveness in bridging digital-physical gaps in component-level construction management, providing a scalable solution for intelligent prefabrication implementation. This methodology advances construction automation by synergizing computational optimization with spatial computing, offering practical insights for industry 4.0 transformation in modular building projects.