Lotsizing

Lot sizing, also called batch sizing, is the process in production planning of determining the quantity of items to produce or purchase in each production run or replenishment cycle. The goal is to balance fixed setup or changeover costs against variable holding costs and stockouts, over a specified planning horizon and subject to capacity constraints.

Inputs include forecasted demand, lead times, setup or changeover costs, unit manufacturing or purchase costs, holding costs, service level targets, and available capacity. The output is a schedule of production quantities and release dates for each period.

Common lot-sizing rules include lot-for-lot (produce exactly the forecasted demand for each period), fixed-quantity (constant batch size), and fixed-interval (produce at regular time intervals to cover demand).

Deterministic models such as the Wagner-Whitin algorithm solve the dynamic lot-sizing problem with known demand by minimizing total costs over the horizon. More complex variants include capacitated lot-sizing problems (CLSP), where capacity limits constrain feasible lot sizes, and multi-item lot sizing with shared resources or substitutions.

Lot sizing is a core component of material requirements planning (MRP) and is implemented in ERP systems to generate order quantities and release dates. The approach affects inventory levels, production scheduling, and service levels.

Benefits include reduced total cost, lower holding or obsolescence risk, and smoother capacity usage. Limitations include dependence on forecast accuracy, potential for suboptimal batching under demand variability, and computational complexity for large, multi-item problems.