Reliability assessment of a stochastic air transport network with late arrivals

•An air transport network with late arrivals and stochastic capacity is studied.•A stochastic-flow air transport network with late arrivals (SATNL) is formulated.•Gain minimal itineraries and maximum tardiness vectors via a DPS-based procedure.•Propose an algorithm to evaluate the reliability as the SATNL’s performance.•Perform impacts of late arrivals, demand and time constraint on the reliability. With the rapid development of low-fare airlines with cheaper fares, the demand for safe and fast air-travel is growing. However, weather shocks, unpredictable situations, or mismatched interactions between related players sometimes impair the on-time landing of air transport networks. This work considers late arrivals and assesses the reliability of an air transport network and its performance indicator in terms of meeting the requested number of passengers under the time constraint. The available number of seats in each flight is regarded as stochastic owing to the number of seats that have been reserved by both travel agents and individual customers. An air transport network is thus formulated as a stochastic-flow air transport network with late arrivals, and all minimal itineraries feasible under the constraint of time are found through a depth-first search based approach. The reliability is derived from all flow vectors such that its capacity and tardiness-related valid conditions are not bounded while fulfilling the given number of passengers. Evaluating the reliability with considering late arrivals implies handling the rapid increase in flights for each tardiness level and facing the influence on the validity of each minimal itinerary. This study also contributes towards potential improvements in the reliability of air transport networks, which can benefit travel agents.