I. Introduction
The competition in airline market [1] goes into a white-hot era, especially, when the cheap flights were involved in the games. The strong needs of airline alliance or flight combinations rise [2]. The flight combination often means the itinerary consists of multiple airports from the original site to the destination site. Although a longer flight in such combination may be required, a cheaper ticket as well as more choices of flight time are offered. There are several advantages in flight combinations. For customers, they can enjoy the cheaper tickets and flexible flight time in their itinerary planning. For airline companies, they can obtain the increase of both full-load ratio and ticket sales in their flights, and thus, increase their revenue. However, the low reliability in some flight combinations may discourage the customer to buy such products [3], and decrease the merits of flight combinations. The low reliability of flight combination means that the connection of flights in some situation becomes impossible, and may cause a customer loss. A study on this problem is required, but there are very limited researches exploring this issue in the literature. This paper proposes a method to evaluate the reliability in flight combination in terms of timed stochastic-flow network.