Automated project rescheduling with preemptive activity splitting for baseline restoration

This paper introduces an automated project scheduling framework tailored to the complexities of construction environments, with a particular emphasis on preemptive activity splitting. The proposed method builds on the preemptive multi-mode resource-constrained project scheduling problem formulation and integrates systems thinking with scenario-based optimization to support dynamic rescheduling decisions. By enabling activities to be interrupted and resumed across alternative modes, the framework better captures real-world project execution dynamics compared to conventional deterministic methods. Weekly baseline plans serve as reference points to guide a rolling-horizon rescheduling process that aims to minimize execution deviations while maintaining cost efficiency. Applied to a mechanical system installation case, the framework achieves a 17.57% reduction in recovery cost relative to conventional compression methods, demonstrating improved adaptability and reduced reliance on extensive input data. These results demonstrate the approach's strong adaptability and its practical potential for construction planning systems, particularly in volatile project environments.