Optimal stochastic inventory-distribution strategy for damaged multi-echelon humanitarian logistics network

The quick distribution of relief goods is vital in alleviating human suffering in affected areas. A relief strat-egy that delivers (pushes) goods to the affected population depending on the predicted demand is crit-ical, especially during the early post-disaster period when accurate demand information is lacking. Sev-eral inventory-distribution strategies based on multi-echelon humanitarian logistics networks have been previously investigated to facilitate a quick response to demand information. However, natural disasters have raised doubts about the performance of these networks. This paper presents the disaster conditions under which conventional multi-echelon networks are conducive to push-mode strategies that deliver relief goods depending on the predicted demand. Specifically, an approximate solution to the optimal inventory-distribution strategy is analytically derived from a dynamic stochastic optimization problem using deterministic approximation and decomposition. Several numerical validations ensure the optimal-ity and practical applicability of the strategy. The approximate strategy identifies disaster conditions that do not contribute to push-mode strategies in conventional multi-echelon networks and proposes an al-ternative network structure.(c) 2023 Elsevier B.V. All rights reserved.