SCSC2003 Abstract S31944

A Simulation Study of the Passenger Check-in System at the Ottawa International Airport

A Simulation Study of the Passenger Check-in System at the Ottawa International Airport

Submitting Author: Prof. Aaron Nsakanda

Abstract:
The growth of demand in the airline industry is creating congestion at many international airports. Ottawa is Canadaís fifth-largest city and one of the fastest growing labor markets. In 2000, 3.4 million passengers used the airport and this was a growth of more than 40% since the last airportís expansion in 1987. The increased passenger flow requires an improvement of service performance, especially for the check-in service system because of limited physical line space.
This research is concerned with a study of the queues and delays in the passenger check-in system at the Ottawa International Airport. Simulation has been applied to gain insights into the relations between parts of the system, the presence of bottlenecks and their causes. We studied the check-in procedure by repeated data collection at the airport and the results of the data analysis define the inputs of the simulation model. These include the current agentsí working schedules, the passengersí arrival pattern distributions, the passengersí service time distributions, the historical flight load factor, the distributions of the types of passengers, and the flight schedule. The scenarios evaluated include changing the queue structure and considering alternate agentsí working schedules. The performance measurement retained are the average waiting time in queue, the maximum waiting time in queue, the average queue length, the maximum queue length, the staff resource utilization
percentage, and the distribution of passengers waiting time in queue.
The critical factor that impacts the check-in service performance proved to be the employee schedule of the check-in agents. A linear programming (LP) model was developed to provide alternate agent working schedules that minimizes the total person hours and meets the passenger loads that vary throughout the day. Since the constraint matrix of the LP model is a totally unimodular (TU) matrix, the optimal solution is always integral. The outputs of this alternative indicated significant improvement in the performance measures discussed above. Sensitivity analysis was performed on changes of passenger loads and service rates. We found that with various passenger loads and service rates, the alternate schedule outperformed the current schedule. Economic class passengersí check-in service performance is more sensitive to the changes of passenger loads than that of business class passengers. This demonstrates that more employees are needed for economic class passengerí check-in du
ring the peak times.
Finally, further considerations regarding simulation model enhancements as well as other operational areas are considered in this presentation.



Back to SCSC2003 Abstracts