Robust optimization for hurricane preparedness

Xinfang Wang

doi:10.1016/j.ijpe.2019.07.037

Return

Robust optimization for hurricane preparedness

2020-03-01

PRE

Xinfang Wang

Jomon Aliyas Paul *

DOI：10.1016/j.ijpe.2019.07.037

Original

Original request for help

Save

Abstract

En 中文

This paper applies the concept of social cost (i.e., logistic, deprivation, and fatality costs) to analyze the optimal deployment time, location and capacities of stockpiles for Points of Distribution (PODs) in hurricane preparedness. We first propose a single-stage, adaptive robust model to determine the optimal deployment time, given the time-variant characteristics of hurricanes. The model is nested in an optimal stopping-time framework that captures the trade-offs between increasing deployment costs and reduced uncertainty as the hurricane approaches landfall. Once the optimal deployment time has been determined, we then propose a less conservative, two-stage robust optimization model with recourse actions to determine the PODs' location, stockpile capacities and flow. Tested on a case study, results show that 1) a non-adaptive model leads to poor decisions about the optimal deployment time; 2) improperly modeled deprivation costs pose significant hidden risks to decisionmakers; 3) deprivation costs increase $3-5 for every dollar cut when the available budget is strictly binding; and 4) significant savings in social cost result from the wait-and-see strategy implemented through the two-stage robust model.

Keywords:

Disaster planning

Location-allocation

Deprivation costs

Adaptive robust optimization

Two-stage robust optimization

Recourse actions

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.