An Agent-Based Simulation-Optimization Approach to Identify Threat Management Strategies for Water Distribution Systems

Threat management decisions in the event of contamination of a water distribution system must be tailored to protect public health, maintain fire-fighting flows and flows to critical care facilities, and avoid inciting panic due to false alarms. Threat management strategies may be composed as a set of rules for taking actions to mitigate the situation and protect public health, such as flushing the water distribution system by opening fire hydrants or inducing demands, isolating portions of the system using control valves, and broadcasting boil water orders. These strategies must be robust to control contamination events that vary in time of injection, duration, and mass flow profiles, as well as to avoid false alarms. The development of threat management strategies can be improved through the use of a simulation-optimization framework that simulates the complex interactions between managers' operation decisions, consumers' water consumption choices and the response of the hydraulics and contaminant transport in the water distribution system through mechanistic and dynamic methods enabled by agent-based models. Heuristic optimization methods are coupled within the dynamic system simulation framework to allow identification of efficient threat management strategies to achieve public health protection and maintain acceptable service. These methods will be explored for an illustrative case study to identify strategies to achieve these objectives.