Design of screening and surveillance strategies for multiple applications.

摘要

Screening and surveillance strategies are widely used in a variety of applications including homeland security, quality control, healthcare, and preventive maintenance. In homeland security, screening strategies are used to identify shipping containers and passengers carrying illegal items. In quality control, screening is aimed at uncovering products defects. In healthcare, surveillance strategies are carried out to monitor patients with conditions that can trigger the development of chronic diseases. In preventive maintenance, inspections are regularly performed to monitor asset condition and prevent failure. In general, screening strategies are procedures comprised of sequences of non-invasive tests used to distinguish unsuspicious from suspicious entities that require more expensive and time consuming manual inspections. The objective is to determine the optimal sequence of tests that maximize detection of suspicious entities, and minimize resources unnecessarily spent screening unsuspicious ones. The design of surveillance strategies serves the same purpose, however, surveillance strategies are carried out over time, and consist of screening strategies performed at scheduled periods. This dissertation presents a series of mathematical frameworks to optimize the design of screening and surveillance strategies considering detection-rate, processing cost, processing time, and diversified screening. The design of screening and surveillance strategies focuses on: (1) selection of screening tests, (2) selection of utilization sequence, and (3) selection of tests settings. Examples are presented to illustrate applications in homeland security and healthcare. In the first case, results are compared with the ones reported in the literature; in the second case, results are validated using ordinal optimization.