A multiobjective model for imperfect sensors layout

In recent years there is a growing concern around the world over the security of water distribution systems. Water distribution systems are spatially diverse and thus are inherently vulnerable to deliberate terrorist contamination intrusions. A common assumption in most sensors optimal layout models for protecting water distribution systems against deliberate intrusions is that sensors are perfect. However, in reality sensors will detect a contaminant only at a specific probability. This study suggests a simple methodology for incorporating imperfect sensors in the decision process of sensors placement. The algorithm explores in a multiobjective framework the detection likelihood of the imperfect sensors versus ideal sensors for different design objectives (e.g. expected time of detection). The methodology is demonstrated using a mid size water distribution system example.