To the best of our knowledge, research in the dynamic plant layout problem (DPLP) assumes that the planning horizon is
fixed and that material flows are known with certainty. But in practice, many companies use rolling planning horizons. Further,
they have to deal with the effect of uncertainty in material flow forecasts. This paper investigates the performance of algorithms
under fixed and rolling horizons, under different shifting costs and flow variability, and under forecast uncertainty. Nearly 1800
problems were run using different algorithms. The results show that algorithms that dominated under fixed horizons may not
work as well under rolling horizons. Also it is difficult to identify an algorithm that performs well under all situations. Thus
the development of efficient and effective heuristics might be useful in solving the rolling horizon problem. It also appears
that increasing the planning horizon under rolling plans does not offer any advantage. Further forecast uncertainty may not
significantly affect the performance of algorithms and in some cases may be beneficial.