This study presents a Mixed-Integer Linear Programming (MILP) model that minimizes the makespan by optimally assigning tasks and workers to stations while accounting for workers' heterogeneous learning rates. Unlike traditional approaches that assume a balanced task distribution, our model leverages individual learning speeds to enhance overall productivity.

Applying the proposed model to a benchmark problem, we demonstrate that when operators have varying learning abilities, an unbalanced task assignment strategy results in a shorter makespan. Specifically, our findings indicate that workload distribution should align with operators’ learning speeds rather than being evenly distributed across stations. Furthermore, we show that conventional strategies, such as random worker allocation or balanced work assignment, can significantly increase the makespan, underscoring the importance of strategic workforce allocation in learning-sensitive production environments. These insights provide valuable guidance for improving assembly line efficiency in small-batch manufacturing, where learning effects are pronounced.