Routing algorithms as tools for integrating social distancing with emergency evacuation

Abstract

One of the lessons from the COVID-19 pandemic is the importance of social distancing, even in challenging circumstances such as pre-hurricane evacuation. To explore the implications of integrating social distancing with evacuation operations, we describe this evacuation process as a Capacitated Vehicle Routing Problem (CVRP) and solve it using a DNN (Deep Neural Network)-based solution (Deep Reinforcement Learning) and a non-DNN solution (Sweep Algorithm). A central question is whether Deep Reinforcement Learning provides sufficient extra routing efficiency to accommodate increased social distancing in a time-constrained evacuation operation. We found that, in comparison to the Sweep Algorithm, Deep Reinforcement Learning can provide decision-makers with more efficient routing. However, the evacuation time saved by Deep Reinforcement Learning does not come close to compensating for the extra time required for social distancing, and its advantage disappears as the emergency vehicle capacity approaches the number of people per household.

Vehicle Routing Optimization for Pandemic Containment: A Systematic Review on Applications and Solution Approaches

“…Other methods: Heuristics, simulation, and machine learning were rarely used in the reviewed articles. However, to provide a good solution for VRPs in real time, a biasedrandomized heuristic with parallel computing was used to generate the solution within milliseconds [92,107], while deep reinforcement learning was implemented in Tsai et al [98] to enhance the routing efficiency. To schedule multiple vehicles simultaneously in a dynamic traffic condition, Mook et al [101] studied the simulation tool of Simio for the purpose of simulating autonomous cars after configuring the routing decision strategy in the road network.…”

“…In addition to DARP model, other articles in this scope of research used the Capacitated VRP model to capture the characteristics of the passenger transportation problem. A WWO (water wave optimization)-based metaheuristic [75] and DNN (Deep Neural Network)-based solution procedure [98] have been applied to implement transportation strategies. The former solution approach has shown success in controlling the pandemic, while the latter can provide a more efficient routing plan than the sweep algorithm.…”

“…Response time was the factor used most often in the social dimension. The minimization of total time in the route was used in many articles [51,89,98,104,106,107]. In order to optimize the bottleneck route, Pacheco and Laguna [91] proposed an alternative objective function, aiming to minimize the longest route time and ignore all shorter routes.…”

“…After contagious transportation, the interior of the ambulances needs to be disinfected before another use [84,88]. Focusing on the example of evacuation, one study [98] suggested integrating social distancing strategy with transport operation-e.g., limiting the numbers of passengers allowed in one vehicle.…”