A New Mathematical Model to Cover Crew Pairing and Rostering Problems Simultaneously

Abstract

Crew scheduling problem includes two separate sub-problems, namely crew pairing and crew rostering problems. Solving these two sub-problems in a sequential order may not lead to an optimal solution. This study includes two main novelties. It combines these two sub-problems and presents them in a single model. On the other hand, despite previous researches that considered a pairing continuously, the proposed model benefits from the capability of considering one or more days off in a paring assigned to a crew member. This is very useful as it enables crew to participate in required courses, doing medical checks etc. Two solution approaches, namely; Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) and are used to solve the model. Eventually, the performance of the proposed algorithms is evaluated. Both ended to satisfactory results; however, PSO relatively outperformed GA in terms of solution optimality and computational time.