Rule-Based AI in Emergency Response Coordination

Abstract

One of the most significant challenges in emergency response management is coordination across different involved agencies (Turoff and Chumer, 2004). An emergency response might rapidly increase in complexity as the number of parties involved increases, and it may become difficult or even impossible for a single individual to make rapid, well-informed coordination decisions. Even seemingly trivial tasks regulated by checklists are prone to faulty coordination decisions during acute stress (Stolpe and Hannay, 2021). This thesis builds on the theory and concept set forth by Stolpe and Hannay, 2021 of how AI planning can be used to solve delegation-and-sequencing problems using Answer Set Programming (ASP). The design science research methodology is used to develop a design concept that illustrates how AI planning can augment emergency response managers’ coordination capabilities in complex, multi-agency incidents. The design concept is evaluated with relevant candidates from Norwegian emergency services, and their feedback forms the basis for the requirements of a technical implementation. The results of the evaluation show that the system could help reduce the burden on the emergency response managers by providing up-to-date incident information and improving the utilisation of external resources.