120 likes | 233 Views
Designing and analyzing an overbooking policy for an airport to determine the cost of compensating overbooked passengers. Utilizing binomial distribution and Monte Carlo simulation to estimate probabilities and simulate passenger numbers. The study considers volunteer compensation, forced passenger costs, and total compensation calculations based on wait times and ticket prices. Recommendations for future work include assessing goodwill costs, altering distributions, and exploring rebooking on external airlines for individualized wait times.
E N D
Cost of Compensation for Overbooked Passengers Kennedy Murphy, Zhongxu Zhao, Alexis Christie
Goals • Design an overbooking policy for an airport • Run simulations • Analyze the potential cost of compensation
Randomly Generate Number of Passengers • Binomial distribution • Probability mass function: P(X=k) = n! pk(1-p)n-k k!(n-k)! p = probability of one passenger showing up for their flight = 0.85 1-p = probability of failure = 0.15 n = overbooking maximum k = number of passengers who show up
Randomly Generate Number of Passengers Ex. For a plane with a maximum capacity of 150 passengers and an overbooking maximum of 177 passengers, the probability that 156 passengers will show up is: P(X=156) = 177 (0.85)156 (1-0.85)177-156 156 = 0.071
Monte Carlo Simulation • Used to simulate values of a discrete random variable • Want to generate a random number of passengers based on binomial distribution http://office.microsoft.com/en-ca/excel-help/introduction-to-monte-carlo-simulation-HA001111893.aspx
Assumptions • Maximum wait time = 24 hr (1440 min) • Everyone that gets kicked off the plane waits the same amount of time for their next flight (i.e. all passengers kicked off the plane are rebooked on the same flight) • # of volunteers = 5 (approx. 3% of flight population) • We always rebook passengers on our own company’s flights • Ticket price = $230 (one way) • Based on Air Canada rate for flight from Halifax to Toronto on a Wednesday at the end of November (i.e. typical day to fly)
Volunteers • U(v) = cost of volunteers • v = # of volunteers We pay each volunteer $200 cash. Thus, • U(v) = 200v
Passengers Forced Off Plane • M(t, f,m) = cost of passengers involuntarily forced off plane • t = wait time • f = fair price (i.e. ticket price) • m = # of people forced off plane • M(t, f,m) = m*00< t <=1 m*min(2f, f + 200) 1< t <=2 m*min(3f, f + 400) 2<=t • M(t, 230,m) = m*0 0< t <=1 m*min(460, 430) 1< t <=2 m*min(690, 630) 2<=t
Total Cost • C(v,m,t,f) = U + M = cost of compensating both volunteers and passengers forced off plane
Future Work • “goodwill” cost • Distribution for volunteers • Altering the wait time distribution • Rebook on outside airlines • Individualized wait times