Resumen de Methodology for optimal design of efficient air transport network in a competitive environment
This thesis aims to dissert about air transport network design taking into consideration the current needs about efficiency in a very competitive industry. The main focus for this work is the airline's point of view and for this reason is going to be common to talk about profitability. A methodology is proposed to analyse current networks and to introduce modifications. First, an analytical approach is proposed with the aim to understand better the interaction of key parameters in network design at strategical level. Analyitical models demonstrate to be very useful because with a few parameters it is possible to decide if point to point or hub and spoke configuration suits better for an airline, given a set of supply conditions. Furthermore, the performance of networks with stopover configuration is tested, basically because for long routes with low demand this newtorks work well, how it was demonstrated by public transport systems. The differenciation of this research, compared against previous studies, is to include the performance of new entrant airlines in the industry. This aspect is carried out evaluating accurately the impact of fixed costs in the evaluation of cost operators. Despite airlines tend to consider aircraft ownership costs and labor costs as variable ones, they are not. They had an important impact in profit and loss account. A bad utilitzation factor of airplane or crew staff is very undesirable for operator and it can consider do not operate new routes if the resources are not well used. Secondly, analytical models are powerful but it is not possible to evaluate accurately the daily operational aspects for real networks with real flight schedules because these models take into consideration average values of some parameters. For this reason, this thesis develops a Tabu search algorithm to carry out airline network planning, based on linear models of each different planning problem. To do this contribution, these models are developed before for linear programming and they are solved with a combination of complete enumeration algorithm and exhaustive search algorithm. Both algorithms provide exact solution or global optimum for any problem statement. Later, Tabu search algorithm improves performance of searching with lower computational cost. The main conclusion is that Tabu developed is a better tool for airline planning than exact techniques because of lower times of computation. This performs better for large networks that, finally, are real networks. However, exact techniques could be interesting for small airlines that can be start-ups. Third, one of main reasons to develop a quantification of costs for managing airlines is due to the problem of complexity in networks. With the growth of air transportation and congestion at airports, primary or reactionary delays become higher day after day. Reactiveness of network depends of configuration and planning. Airlines manages it with extra reources, one way is allocating buffer times in flight schedule and other is having aircrafts at ground at main bases to enter in service and recover flight plans. Airline network design, planning and efficient algorithms are key assets to provide robustness for airlines. Finally, driving airlines in high competitive environments are difficult. Running low costs is the main decision for managers. However, analysing the problem from theory game point of view, allows finding key reasons to support this argument. First, a Stakelberg model is defined for two competing airlines. This model demonstrates that a war on frequencies or fares damages both competitors. Furthermore, a Cournot model it is shown and it proposes a navigation fee attending to correct utilization of capacity. Both models are a theoretical framework but demonstrate consistency and encourage further investigations.
