An Improved Genetic Algorithm for Crew Pairing Optimization

Abstract

Crew pairing is a sequence of flights beginning and ending at the same crewbase. Crew pairing planning is one of the primary processes in airline crew scheduling; it is also the primary cost-determining phase in airline crew scheduling. Optimizing crew pairings in an airline timetable helps minimize operational crew costs and maximize crew utilization. There are numerous restrictions that must be considered and just as many regulations that must be satisfied in crew pairing generation. The most important regulations—and the ones that make crew pairing planning a highly constrained optimization problem—are the the limits of the flight and the duty periods. Keeping these restrictions and regulations in mind, the main goal of the optimization is the generation of low cost sets of valid crew pairings which cover all flights in the airline’s timetable. For this research study, We examined studies about crew pairing optimization and used these previously existing methods of crew pairing to develop a new solution of the crew pairing problem using genetic algorithms. As part of the study we created a new genetic operator—called perturbation operator.Unlike traditional genetic algorithm implementations, this new perturbation operator provides much more stable results, an obvious increase in the convergence rate, and takes into account the existence of multiple crewbases.