Assembly scheduling subject to outsourced-component available time constraints

Abstract

Two-stage assembly flowshop scheduling problems are investigated in this thesis. In a general two-stage assembly flowshop, to complete each job, multiple components are fabricated by in-house machines or procured from outside suppliers at the first stage. After all components of each job are prepared at the first stage, the components are assembled at the second stage. When the available times of outsourced-components can be completely adjusted in favorable to assembly schedule, it is not necessary to consider the outsourcing components any more. However, if not, then the fabrications of in-house components and the subsequent final-assembly should be scheduled after due consideration of the arrival times of outsourced-components. Two types of two-stage assembly scheduling problems are considered in this thesis. The first type of problem has an outsourced-component whose available time should be considered as constraint for whole scheduling. In the problem, each job is assembled with two types of components for which makespan is the objective measure. An in-house dedicated-machine fabricates one type of component and the other type of component is outsourced subject to its own arrival time, which should be considered as constraint for whole scheduling. It is proved that the problem is NP-hard in the strong sense and some solution properties are characterized. Based on the solution properties, three heuristic algorithms and a branch-and-bound algorithm are derived. Through numerical experiments, it is shown that the performances of the derived algorithms are quite good. For the same problem, four ρ-approximation algorithms and their tight worst-case error bounds are derived. Moreover, one of them is modified to give an improved worst-case performance ratio of 3/2. It is also proved that the worst-case performance ratio is best possible. The other type of problem is related with the case in which there is not any component available time constraint. The probl...