ID:
publications-4418
Type:
article
Year:
2010
Authors:
Xu, Jianhua and Small, Mitchell J. and Fischbeck, Paul S. and VanBriesen, Jeanne M.
Title:
Integrating Location Models with Bayesian Analysis to Inform Decision Making
Venue/Journal:
Journal of Water Resources Planning and Management
DOI:
10.1061/(asce)wr.1943-5452.0000026
Research type:
Water System:
Technical Focus:
Abstract:
In the present work, we locate sensors in water distribution networks and make inferences on the presence of contamination events based on sensor signals. We fully consider the imperfection of sensors, which means that sensors do provide false positive and false negative signals, and we propose a two-stage model by combining a facility location model with Bayesian networks to (1) identify optimal sensors locations and (2) infer the probability of the occurrence of a contamination event and the possible contamination source based on sensor signals, the probability of a contamination event being detected by the sensors given that there is a contamination event, and the probability of detecting a contamination event given that there is actually no such event (overall false positive rate). This two-stage model can also be used to construct the trade-offs between the number of sensors and the power (the false negative and false positive rates) of individual sensors while guaranteeing the performance (the proba...
Link with Projects:
Link with Tools:
Related policies:
ID:
