Bi-Objective Master Production Scheduling Considering Production Smoothing: A Case Study in the Truck Industry

In the context of mass customization and mixed-model production systems, Master Production Scheduling (MPS), which determines production start dates, plays a critical role. However, in such environments, MPS faces a dual challenge: ensuring due-date adherence under multiple capacity constraints while also reducing operational instability caused by uneven day-to-day consumption of critical components, referred to as Replenishment and Industrial Characteristics (RICs). This paper proposes a new mathematical model for MPS with a Smoothing Mechanism for RICs (MPS-SM). This bi-objective formulation extends a baseline due-date-driven model with an explicit production smoothing/leveling (also known as Heijunka) term, minimizing deviations of RIC usage from weekly ideal levels. By embedding smoothing directly into MPS, the approach provides a pre-leveling effect that can reduce (or ideally eliminate) downstream complexity, specifically related to schedule modifications required in a separate smoothing stage. To reflect changing scheduling priorities, smoothing is weighted through an innovative context-aware non-linear weekly function that assigns lower importance near execution and greater importance farther into the horizon. The models are evaluated in a rolling-horizon simulation-optimization framework using data from a real-world truck manufacturer. Several experiments over 300 discrete-event simulated days show that MPS-SM consistently reduces RIC variability while inducing a controlled increase in lateness penalties.