TY - JOUR
T1 - An evolutionary algorithm for port-of-entry security optimization considering sensor thresholds
AU - Concho, Ana Lisbeth
AU - Ramirez-Marquez, Jose Emmanuel
PY - 2010/3
Y1 - 2010/3
N2 - According to the US Customs and Border Protection (CBP), the number of offloaded ship cargo containers arriving at US seaports each year amounts to more than 11 million. The costs of locating an undetonated terrorist weapon at one US port, or even worst, the cost caused by a detonated weapon of mass destruction, would amount to billions of dollars. These costs do not yet account for the devastating consequences that it would cause in the ability to keep the supply chain operating and the sociological and psychological effects. As such, this paper is concerned with developing a container inspection strategy that minimizes the total cost of inspection while maintaining a user specified detection rate for "suspicious" containers. In this respect and based on a general decision-tree model, this paper presents a holistic evolutionary algorithm for finding the following: (1) optimal threshold values for every sensor and (2) the optimal configuration of the inspection strategy. The algorithm is under the assumption that different sensors with different reliability and cost characteristics can be used. Testing and experimentation show the proposed approach consistently finds high quality solutions in a reduced computational time.
AB - According to the US Customs and Border Protection (CBP), the number of offloaded ship cargo containers arriving at US seaports each year amounts to more than 11 million. The costs of locating an undetonated terrorist weapon at one US port, or even worst, the cost caused by a detonated weapon of mass destruction, would amount to billions of dollars. These costs do not yet account for the devastating consequences that it would cause in the ability to keep the supply chain operating and the sociological and psychological effects. As such, this paper is concerned with developing a container inspection strategy that minimizes the total cost of inspection while maintaining a user specified detection rate for "suspicious" containers. In this respect and based on a general decision-tree model, this paper presents a holistic evolutionary algorithm for finding the following: (1) optimal threshold values for every sensor and (2) the optimal configuration of the inspection strategy. The algorithm is under the assumption that different sensors with different reliability and cost characteristics can be used. Testing and experimentation show the proposed approach consistently finds high quality solutions in a reduced computational time.
KW - Container inspection
KW - Continuous evolutionary optimization
KW - Decision-tree
KW - Sensor thresholds
UR - http://www.scopus.com/inward/record.url?scp=73649122395&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=73649122395&partnerID=8YFLogxK
U2 - 10.1016/j.ress.2009.10.006
DO - 10.1016/j.ress.2009.10.006
M3 - Article
AN - SCOPUS:73649122395
SN - 0951-8320
VL - 95
SP - 255
EP - 266
JO - Reliability Engineering and System Safety
JF - Reliability Engineering and System Safety
IS - 3
ER -