A goal programming model for early evacuation of vulnerable people and relief distribution during a wildfire

Abstract

Wildfires generate many safety issues to population at risk areas. Sometimes, due to the proximity of a fire, it is necessary to carry out the evacuation of vulnerable people in order to protect them from heat, sparks or flames. To facilitate the evacuation planning, potential evacuees are classified according to their health condition, so that they are taken care of properly. Additionally, once they reach a safe area, basic supplies are crucial to ensure their welfare. In order to assure the proper coverage of the basic needs of the evacuees, supplies have been also classified according to their characteristics. Furthermore, not everybody decides to evacuate at the same time; on the contrary, vulnerable people affected by fires require assistance or go to the designated pick-up points according to their own susceptibility about the situation, leading to the dynamic arrival of potential evacuees along a certain time horizon. In the same way, dynamic arrival of potential supplies along time can be considered. All these factors define a real-life problem that can be formulated as a mixed integer optimization model that is dynamic, multi-modal and multicriteria. This model is more flexible than others available in the literature and allows for a joint approach to people evacuation and supply distribution. A case study regarding the Saddleridge Fire that hit San Fernando Valley in Los Angeles County, California, in October 2019, is introduced and used to validate the proposed model and evaluate its performance.