Vehicle routing problem for humanitarian relief distribution under hybrid uncertainty

Abstract

PurposeGiven the importance of estimating the demand for relief items in earthquake disaster, this research studies the complex nature of demand uncertainty in a vehicle routing problem in order to distribute first aid relief items in the post disaster phase, where routes are subject to disruption.Design/methodology/approachTo cope with such kind of uncertainty, the demand rate of relief items is considered as a random fuzzy variable and a robust scenario-based possibilistic-stochastic programming model is elaborated. The results are presented and reported on a real case study of earthquake, along with sensitivity analysis through some important parameters.FindingsThe results show that the demand satisfaction level in the proposed model is significantly higher than the traditional scenario-based stochastic programming model.Originality/valueIn reality, in the occurrence of a disaster, demand rate has a mixture nature of objective and subjective and should be represented through possibility and probability theories simultaneously. But so far, in studies related to this domain, demand parameter is not considered in hybrid uncertainty. The worth of considering hybrid uncertainty in this study is clarified by supplementing the contribution with presenting a robust possibilistic programming approach and disruption assumption on roads.

Designing a post-disaster humanitarian supply chain using machine learning and multi-criteria decision-making techniques

“…Immediate response is one of the most relevant issues in the HSC. With the proper redistribution of resources, an optimal allocation can be achieved in these critical situations (Nodoust et al, 2021). With immediate response in mind, and intending to minimize the overall consumption of relief operations and the optimal allocation of resources, Sarma et al (2020a) presented a model considered in a mixed uncertain environment.…”

“…This highlights the managerial challenges in the humanitarian supply chain and the need for prior preparation for an appropriate response. In addition, the high sensitivity of demand to natural disasters over time poses the challenge of meeting demand in the short term (Cai et al, 2020;Chen, 2022;Nodoust et al, 2021;Zhang et al, 2023). In a situation where rapid change in demand, high time-sensitivity of demand and lack of resources occur, managers in the humanitarian supply chain will urgently need to consider and use planning to be highly prepared for relief operations.…”

“…In addition, the high sensitivity of demand to natural disasters over time poses the challenge of meeting demand in the short term (Cai et al. , 2020; Chen, 2022; Nodoust et al ., 2021; Zhang et al. , 2023).…”

A Multi-Depot Dynamic Vehicle Routing Problem with Stochastic Road Capacity: An MDP Model and Dynamic Policy for Post-Decision State Rollout Algorithm in Reinforcement Learning

“…On the other hand, Ref. [22] presented a two-stage Location Routing Problem (LRP) of distributing first aid relief materials post-earthquake disaster where the complex demand uncertainty is addressed. Here the mixture of uncertain demand from the perspective of randomness (probabilistic theory) and fuzziness (possibility theory) due to the merging of subjective and objective data forms the hybrid demand uncertainties.…”

“…Even with PDS introduced as above, the computation for an optimal decision is usually challenging, especially when computing the value of the PDS, (22). The ADP approach approximates the value of PDS instead.…”