Comparing two aggregate planning models

The sensitivity of the linear programming (LP) model to forecast errors in aggregate planning (AP) is evaluated and its performance compared with that of the production switching heuristic (PSH) model. Two versions of the PSH, the popular three-level model and a two-level extension, are considered. While the LP tends to outperform the PSH in situations with high forecast errors, the performances of the models are comparable for low to moderate forecast errors. The use of an additional production and workforce level in the PSH is not justified unless forecast errors are high. A variation in the forecast error distribution had no significant effect on these findings.