A Robust Multi-commodity Rebalancing Process in Humanitarian Logistics

Abstract

After disasters occurred, many refugees have to suffer a lot. To relieve this detrimental situation, various commodities are distributed to the pre-determined warehouses. However, the initial multi-commodity distribution may be imperfect, which results in some warehouses having surplus commodities compared to other unmet warehouses. Hence, it is necessary to rebalance commodities among those warehouses. Because of the uncertain environment after a disaster, the demand is usually uncertain. To plan this multi-commodity rebalancing process appropriately, it is usually assumed that the collected data are uncontaminated. However, this assumption can be easily violated due to the uncertain environment or human error, which results in the biased estimation of the solution. In this study, we propose a strategy for remedying the difficulties associated with data contamination so that a set of robust decisions are obtained. Through a case study, we show that the proposed strategy facilitates effective decision-making in the multi-commodity rebalancing when the data contamination is involved.

Multi-modal transportation planning for multi-commodity rebalancing under uncertainty in humanitarian logistics

“…Due to the dynamic environment after a large-scale disaster, the collected information is usually uncertain (Kostoulas et al 2008;Haghi et al 2017;Gao and Jin 2020;Balcik and Yanıkoğlu 2020). Various uncertain elements need to be considered to improve the reliability of studies in humanitarian logistics.…”

A note on the warehouse location problem with data contamination

“…The facility location problem is an important issue in supporting the physical distribution as it contributes significantly to the travel time or cost in logistics systems [9,20,34,39]. Because the travel time or cost can be analyzed by discrete and continuous aspects in the space, the facility location problem is commonly divided into two categories; namely, discrete-space and continuous-space facility location problems.…”

“…Another possible reason for the existence of an outlier is that the underlying distribution of the sample has fat tails so that the observation is not contaminated but just an extreme observation from one of the heavy-tailed ends of the underlying distribution. For more details on the outliers and their effect on the estimator behavior, one should refer to Tukey [67], Rousseeuw and Croux [63], Park [60] and Gao and Jin [34]. This data contamination and variability in parametric value have been receiving insufficient attention in disaster response.…”