The impact of human consideration,schedule types and product mix on scheduling objectives for unpaced mixed-model assembly lines

There is an increasing awareness in scheduling research that human behaviour needs to be considered explicitly in scheduling models. Although most scheduling literature ignores human behaviour, especially sequence-dependent processing times form a good basis for explicit consideration. Hence, a processing time function is derived that considers the effects of learning, forgetting, fatigue and recovery. The necessity for explicit human consideration can be regarded as most urgent for unpaced highly-manual mixed-model assembly lines. Based on real data a simulation study is conducted to determine the effect of explicit human consideration while also taking into account the effects of different idealised schedule types and the product mix. The results strongly indicate that the product mix has a consistently high impact on scheduling objectives, the schedule type affects lower-level objectives like starving and blocking times to a greater extent than higher-level objectives like makespan and flow time, and that for certain objectives the height of the objective values and the relative favourability of schedule types depends on human consideration.     

Modelling,analysis and improvement of an integrated chance-constrained model for level of repair analysis and spare parts supply control

In traditional practices, maintenance system and spare parts inventory control are usually considered in isolation, resulting in suboptimality. In a military system, the level of repair analysis (LORA) is often employed to help operate its repair networks. In this paper, we consider an integrated LORA and inventory control problem and formulate this problem as a mixed-integer nonlinear programming problem with chance constraints. Two second-order cone constraints are proposed to approximate the chance constraints. Furthermore, we propose an outer approximation (OA) algorithm based on the OA cuts. Extensive numerical results show that the OA algorithm significantly improves the computational efficiency under various types of components and network complexity. Next, we investigate the influence of service level and resource capacity, and propose the findings. Our results indicate that a higher service level leads to steeper costs, more resources, larger storage and heavier repair burdens at operating sites. Moreover, enhancements in resource capacity from the status quo lead to improvements in repairs and shrinkage in discards, bringing direct economic benefits. The insights extend to uncertain settings. It may be initially counterintuitive for many practitioners that demand uncertainty poses relatively subtle impacts.