AGGREGATE PLANNING IN HYBRID FLOWSHOPS

We present a linear programming based heuristic for the solution of a class of aggregate level planning problems in hybrid flowshops (flowsh ops with several machines per stage). First, the general planning prob lem is modelled as multi-level with parallel processors, multi-item, c apacitated, lot-sizing with set up times. We suggest a hierarchical ap proach which sequentially loads the stages; each stage is constrained by the solution of its preceding stage and each stage is treated as a multi-item, capacitated, lot-sizing problem with setup times on parall el processors. We show how this latter problem may be reformulated and solved heuristically as a sequence of network problems (trans-shipmen t problems) in which the amount of capacity lost in setups is fixed fo r each period and each processor. The model is within the computing re ach of a PC.