Matthias Viehmann

Abstract

Fleet assignment is a major decision problem of the airline planning process. It assigns fleet types to scheduled flights so that profit is maximized. Basic fleet assignment models are solved irrespective of other planning steps. Although this leads to an optimal assignment solution, it might cause suboptimal solutions in subsequent planning steps. This thesis provides at first a comprehensive overview of existing models integrating fleet assignment and at least one further planning step.

The majority of fleet assignment models do not revise their solution if demand and thus profitability of flights turns out to be different than expected. This thesis presents a detailed analysis of dynamic fleet assignment which combines an initial fleet assignment and later demand-driven re-fleeting. Further, we incorporate robustness into the initial fleet assignment, which we assume to increase re-fleeting flexibility. Additionally, we propose an extension of an existing robustness concept. In computational experiments a dynamic fleet assignment process incorporating robustness is simulated based on a schedule containing 700 flights. The results demonstrate that a robust initial fleet assignment leads to higher re-fleeting profits than a basic fleet assignment model if demand variation is large. Further, our proposed robustness strategy leads to significantly lower run times and achieves greater profits than the traditional robustness concept.