Abstract:
An important class of problems that arise in aviation security is the screening of passengers prior to boarding an aircraft. Developing strategies to effectively and efficiently screen passengers, as well as to allocate and operate screening devices, can be quite challenging. Moreover, even after such systems are in place, it can be very difficult to measure their effectiveness.
The speaker will present an approach for designing a passenger and carry-on baggage-screening system using discrete optimization and sequential stochastic assignment theory. The approach simultaneously addresses two key problems in passenger screening: the device allocation problem (i.e., the purchase and installation of security devices) and the passenger assignment problem (i.e., the operation of security devices). Obtained models provide insights into the operation and performance of aviation security systems, under the assumption that a passenger prescreening system (such as Computer-Aided Passenger Prescreening System) has been implemented and is highly effective in identifying passenger risk.
Then, a general extension to the theory of sequential stochastic assignment will be presented for the case where available resources need to be optimally allocated to a random number of sequentially arriving jobs with stochastic parameters. Sequential and dynamic resource allocation under uncertainty is a common problem in investment planning, transportation, and scheduling. The ability to use historical data to predict future demand is critical in making profitable decisions in these environments. The speaker will address a question of how non-standard discrete distributions, based on such data, can be used to make decisions that maximize the expected profit. The derived closed-form optimal policies will allow managers to use the inside knowledge of the behavior of their customers, and optimize their strategy to guarantee the most efficient use of available resources under well-defined circumstances. The results reported can be applied in the domains of logistics, management, and have direct application to aviation security models to be described in the first part of the seminar.