Producing prefabricated bathroom units (PBUs) involves a dry installation method that heavily relies on human labor. Therefore, strategically accomplishing efficient and agile workforce planning emphasizes the critical significance. However, previous studies that address workforce planning often overlook the long-term stochastic effects and assume a homogeneous workforce for the sake of computational simplicity. To overcome these limitations, this study adopts the Markovian approach to establish an explicit relationship between workforce cost and cycle time, considering a heterogeneous workforce and uncertain human-caused events. The proposed model has a hierarchical structure that addresses the behavioral tendencies that drive task allocation at the individual level and the aggregate effect of manpower allocation at the operation level. By integrating Little's law in queuing theory and metaheuristics optimization, the cycle time can be calculated while searching for the optimal workforce arrangement configuration. The study cross-validated the computational results with empirical data from a precast factory in Singapore and conducted a sensitivity analysis to verify the reliability. The results show that cross-training workers to multiple skills can lead to significant time savings, with a maximum of roughly 23 h saved in PBU cycle time. Ultimately, this research contributes to the body of knowledge by proposing a strategic workforce planning model that accounts for a heterogeneous workforce and uncertain human-caused events. It utilizes a partial cross-training configuration strategy to maximize productivity and flexibility for PBU production. The fit-out process of prefabricated bathroom units (PBUs) relies heavily on manual labor, becoming a bottleneck in the supply chain. Efficient and agile workforce planning is critical to overcome this challenge. However, planning for a large, heterogeneous workforce for tasks in an uncertain environment is difficult. Therefore, this research aims to establish a clear relationship between workforce cost and cycle time to assist with quick decision making during the planning stage. The proposed model has a two-level hierarchical structure that addresses individual-level behavioral tendencies and the operational-level aggregate effect of manpower allocation. This approach ensures optimal workforce size, composition, cost, and flexibility for the project's needs. The model was tested using a real-world industrial case study, and the results showed that cross-training workers in multiple skills can significantly reduce PBU cycle time. The increased workforce flexibility is the main factor responsible for the time savings, with a maximum of roughly 23 h saved.
