A novel constraint propagation approach for airline schedule design problem

Abstract

In this paper, we propose a novel finite domain constraint programming formulation using constraint propagation (CP) approach with variable and value ordering for solving flight scheduling problem. We subsequently model the same problem using time-space network representation and integer linear programming (ILP) technique. The performance of the CP and ILP models are evaluated on simulated and a real life case. The ILP and CP models are solved using CPLEX 12.2 solver and CP optimiser 12.2, respectively. We observe that the ILP model generates better first feasible solutions, while the CP model produces faster feasible solutions. In few cases, the CP model takes longer time compared to the ILP model. For the case data with 600 aircrafts and 1,000 markets, the CP model generates feasible solution rather faster, while the ILP model could not find feasible solution in a day for defined time-space network flow problem with some additional constraints.