The impact of conditional dependence on checked baggage screening

Checked baggage screening, an important layer of an aviation security system, has gained a lot of research interest in the literature. Due to the different characteristics (for example, cost, capacity, and efficiency) of various screening devices, several cost-benefit models have been proposed to help select screening devices to include in a checked baggage screening system. Almost all cost-benefit models involving multiple devices implicitly assume that, given that a bag carries a threat or not, the responses from different screening devices are conditionally independent. This assumption may not hold in real situations because one device’s response may be positively or negatively correlated with another device’s response. In this paper, we bridge this research gap and address the impact of conditional dependence on checked baggage screening based on a more general system configuration and an existing cost-benefit model. The cost-benefit model investigates a two-device screening system with the objective of minimizing the expected cost per bag. Our detailed numerical results suggest that (1) the higher the correlation coefficient, the higher the optimal expected cost per bag, (2) incorporating correlation coefficients may even alter the choice of devices to include in the optimal screening system, and (3) the more the responses are correlated, the higher the relative increase of the optimal expected cost per bag when ignoring conditional dependence.