Dissertation. - Universitaet Hamburg, 2012. - 359 pagesThe research contained in this thesis was undertaken partly as an external doctoral candidate and partly as a research and teaching assistant at the Institute of Information Systems, University of Hamburg. It contains eight articles and a technical report in the field of aggregate production planning and supply chain management. The research question immanent to this
work is how lead times that are load dependent are taken into account in mathematical models for the tactical planning level, how they influence other planning levels and the resulting production plan. There is empirical evidence that lead times exponentially increase with the increase of capacity utilization measured in workload or work-in-process in the production system long before the capacity limit is reached. This can lead to significant differences
in planned and realized lead times. Abstraction from such nonlinearities frequently takes place, mainly in favor of complexity reduction of mathematical models for use in practice and implementation into standard software such as advanced planning systems or enterprise resource planning. Variabilities of lead times may become significant when targeting 100% resource utilization. This leads to longer waiting times of production parts and products and lead to quality losses of such items. In the worst case, they cannot be used for their original purpose, anymore, and have to be discarded or reworked if technically possible. Therefore, another research question is how quality losses and lifetime restrictions are taken into account
in the literature up to date. The latter mentioned aspects pertain to the highly dynamic research field of green supply chain management that contains, among others, questions of rendering production processes more environmentally friendly. This ranges from product design that influences all fields of production to optimal recycling, remanufacturing, and rework processes. It further comprises actions from wastage or disposal reductions also in terms of energy usage translated in related costs that not only imply CO2 emission reductions, but also long-term recovery actions of exploited lands and environments. Related literature surveys show that there do not exist mathematical formulations taking into account all mentioned aspects. We closed this gap by developing discrete dynamic models formulated
as mixed integer programming (MIPs) including production smoothing that accounts for load-dependent lead times (LDLT) thus aiming at avoiding peaks of resource utilization together with lifetime constraints of items as well as rework of items that pass their useful lifetime. Moreover, remanufacturing of externally returned items is integrated together with waiting-dependent rework processing times.
