Rolling horizon heuristics for production planning and set-up scheduling with backlogs and error-prone demand forecasts

Three families of models and fast heuristic methods are developed for identifying a production plan and immediate set-up schedule for a manufacturing line with changeover times. The initial model is exact, but is optimally solvable only for very short planning horizons. A second modelling and solution method optimizes production one period at a time, first in a backward pass to identify target inventory levels, and then in a forward pass to build up these target inventories. Finally, a third method plans set-ups and lots on a period-by-period basis, estimating the capacity usage of future set-ups. All three methods are first tested under static conditions and then on a rolling horizon basis with differing degrees of demand forecast accuracy, tightness of capacity and length of horizon. Computational experiments confirm that even under great forecasting uncertainty the planning horizon should extend beyond the time at which the horizon is rolled forward and the forecasts updated. Tests also show that the degree of capacity tightness and horizon length affects which approximate models and methods are most successful. The degree of forecast error appears to have limited impact on the planning horizon to be used and relative performance of the models.