Formalizing Construction Sequencing Knowledge and Mining Company-Specific Best Practices from Past Project Schedules

In this paper, we present a machine-learning based method that allows company-specific construction knowledge to be automatically learned from past project schedules and weekly work plans without the need for any manual human input. The proposed model is built using long short-term memory recurrent neural networks (LSTM-RNNs) and is trained on construction sequences extracted from previous project schedules. While training, the model learns the likelihoods of different successor alternatives given a sequence of previous schedule activities. Experimental results on 12 real-world schedules show accurate and consistent predictions of potential future activities at various stages of construction. Results also demonstrate the method's ability to formalize sequencing logic and mine what we call dynamic means and methods templates (DMMTs) from previous projects. When used as the engine for a project controls system, this solution has potential to automatically generate schedules using work templates; validate the correctness in the logic of an existing schedule; and revise look-ahead schedules.