An Optimization Approach for Network Structure Design Problems in Short Haul Air Transportation\ud

An attempt has been made in this report to study network formulation and optimization techniques for simple airline networks that could be progressively expanded to have further study at a micro level to accommodate more constraints. Literature review indicates that cost, revenue, and profit models are used for this purpose with the help of LP, ILP, heuristic approaches, and hybrid problems. The air network in India linking five cities is taken as a case study to formulate a simple network problem, using Operating Cost models, with an indirect input of revenue based on traffic variation. The problem is solved using the LP codes available in MATLAB™ optimization toolbox, which produce fractional flight frequencies; these are rounded off to the nearest integer values. Branch and bound heuristic algorithms leading to an integer linear programming approach have also been attempted. A fraction of passenger demand is spilled over for lower levels of aircraft utilization and load factor and peak flight frequency requirements are simultaneously imposed. The results indicate that these constraints contribute more significantly to the spill over effect than the rounding of frequencies to integer. A comparison of the results of some existing models obtained during literature survey for maximum level of service (LOS) when applied to same conditions over different routes or links is presented. Another two models that maximize revenue and profit, respectively, have also been attempted and the results obtained are collated. Nine different cases are applied to these models, beginning with a simple baseline case and then successively increasing the complexity. The capabilities of MATLABTM to formulate the heuristic integer LP problems are also explored in this study. The current study on basic airline network problem and the sensitivity of results indicate that the optimization tools can be further explored to formulate as IP problem to effect the spillover in passenger demand, which will be taken up as a future work