Optimization of container inspection strategy via a genetic algorithm

It is estimated that 90% of the world's freight is moved as containerized cargo, with over 125 million TEUs (Twenty foot Equivalent Units) of container being shipped by 2010. To inspect this volume of cargo for explosives, drugs or other contraband is a daunting challenge. This paper presents an optimization technique for developing an inspection strategy that will provide a specified detection rate for containers containing contraband at a minimum cost. Nested genetic algorithms are employed to optimize the topology of an inspection strategy decision tree, the placement of sensors on the tree and the sensor thresholds which partition suspicious containers (containers believed to contain contraband) from innocuous containers (containers which are believed to be free of contraband). The results of this optimization technique are compared to previously published techniques.