A Bid Prepared for Epsilon Airlines: Minimization of Cost for Transfer Escorts in an Airport Terminal

We address the problem of minimizing the cost incumbent upon Epsilon Airlines to maintain a wheelchair escort service for transfers in an airport terminal. To do this we carefully develop several models which simulate the conditions of an airport terminal to which such an escort service must respond: the geometry of a general airport layout, the probabilistic flow of flight traffic in and out of terminal gates, a probabilistic computation using a binomial distribution for the number of passengers on a given flight who require wheelchair transfer service, as well as a randomized model for transfer destinations within the terminal. Given these dynamic descriptions of terminals and flights, we proceed to develop a rigorous economic model to quantify both the short and long term costs of operating such a service. Sources of this cost include the salaries of escorts, the maintenance and storage of wheelchairs in the terminal, and, most significantly, the costs incurred when late escorted transfers delay a departing flight. The cost of this delay is twofold: There is a cost associated with every additional minute of a plane’s operating while delayed, as well as a long term cost due to the reduction in ticket sales that results from passengers who avoid airlines with increasing delay times. We develop a Simulated Annealing (SA) algorithm that uses our economic models to minimize cost by optimizing the number and allocation of escorts to passengers requiring wheelchair-aided transfers. Having indexed the space of all possible escort allocations in a way that makes it accessible to SA, our algorithm selectively searches the entire space of allocations for a global optimum. Although the space of possible escort allocations is too computationally large to allow for comparison of our algorithm’s results to a globally optimal allocation, our simulations suggest that SA is highly effective at approximating this optimum. Using current airport and airline data, we break our analysis down to short and long term costs, simulating escort service operation under dynamic airport conditions, varying air traffic, airport size, and the fraction of traveling population that requests wheelchair-aided transfer (simulating a greater future abundance of elderly travelers). Page 1 out of 50 Control #: 916