How hospitals response to disasters; a conceptual deep reinforcement learning approach

Abstract

Abstract During a disaster the requests for using ambulance services increases. Efficient assignment of the ambulances leads to lowering the patients' travel time. Simulating these environments is very complex and needs a solid framework. This paper uses a Deep Reinforcement Learning approach to better schedule ambulance dispatch problem during those disasters. The concept of a call and assignment of ambulances are illustrated and the elements of states, rewards, and actions in the formulations are described. The algorithm steps for solving this problem are also presented. This paper can help disaster planners to have a better idea for better scheduling ambulances.