Inventory control and scheduling problems in a single-machine multi-item system

Abstract: This Doctoral Thesis addresses the topic of inventory control and scheduling in a single-machine multi-item system. Specifically, it considers a group of items processed, one at a time, on a single facility. Single-machine multi-item systems occur frequently in practice and apply both in continuous flow processes and batch flow processes. For instance, areas of applicability could include metal stamping, bottling, paper production, food processing, plastic extrusion, printing, and chemical batch production among others. In these cases, it is common to use cyclic schedules for the processing of items. The thesis contains an introductory part and five papers. Two papers present heuristics for determination of cyclic schedules assuming sequence- dependent setups. First, a reverse logistics system, where used products are disassembled, is considered. In this case, setup costs are assumed to be directly proportional to setup times. The heuristic results in disassembly frequencies, idle time, and the sequence in which the items should be processed. The second paper considers production settings and assumes setup costs not directly proportional to setup times. The heuristic presented in that paper also results in frequencies, idle time, and the sequence in which to process the items. These two papers assume deterministic environments. The remaining three papers consider stochastic environments and present planning and control models to be applied under these circumstances. One paper applies deterministic lot sizing models to stationary stochastic demands in a simulation study. A control model is also developed in the paper in order to make the decision for which item to produce and when to produce it. The remaining two papers present planning and control models for determination of safety stocks and order-up-to levels when items are produced in a fixed cyclic schedule. The models can be applied in environments with stochastic demands, stochastic operation times, and stochastic setup times or combinations thereof. The papers in this thesis can be combined in different ways and hence cover a variety of industries and practical applications. Practitioners in the area of production and inventory control would then get models for planning and controlling the processing of multiple items on a single facility. The models are preferably implemented in computerized Enterprise Resource Planning (ERP)-systems at manufacturing companies.