Sociotechnical simulation and evolutionary algorithm optimization for routing siren vehicles in a water distribution contamination event

• Authors: M. Ehsan Shafiee;Emily M. Zechman
• Venue: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation
• Year: 2011

Water distribution contamination incidents occur when a poisonous chemical or pathogen is introduced intentionally or accidentally to the pipe network that delivers potable water to the residents of a municipality. These events pose a challenge to decision makers, who should quickly identify a threat and the most effective response actions for protection of public health. In these events, the dynamic interactions among consumers, utility managers, public health officials, and the water distribution pipe network affect the emergent exposure of consumers. An Agent-Based Modeling (ABM) approach is used to simulate the interactions among agents and flow conditions in the water distribution system to provide an understanding of effects of dynamic and adaptive behaviors on public health. While utility operators can protect consumers using a wide range of protective and mitigative responses, routing of siren vehicles can be effective as consumers are warned about a contaminant in the water system and respond by stopping different water activities, such as drinking water. Development of crisis management routing strategies, which are a set of routes to best warn and protect consumers from exposure, is enabled through a new simulation-optimization framework. A genetic algorithm and the ABM are coupled to find routes for siren vehicles that minimize the number of consumers who are exposed to contaminated tap water. The framework is demonstrated for an illustrative case study, a mid-sized virtual city, to identify efficient routes for protecting public health.