Robust Optimization of Uncertain Logistics Networks

Decision makers should concentrate on managing any probable risk of the logistics system, starting from the design phase. As a result, unforeseen conditions during implementation will be less likely to invalidate the basic design plan or disturb the performance targets. One of the main difficulties of the logistics management problems is how to wisely consider the uncertainty about the future in the modeling phase. In the real world, the existence of noisy, incomplete, or erroneous information and data is an unavoidable fact that widely affects the efficiency of the logistics network processes at the implementation phase, so not correctly modeling for these inherent uncertainties might result in impractical plans. Hence, one critical role of logistic managers relates to the way of facing noisy and uncertain environments in order to obtain more effective networks with less re-planning. The importance of this subject has caused a considerable growth in the number of studies dedicated to the uncertainty in supply-chain and logistic networks and their associated modeling approaches to optimize the design and performance of the networks under uncertainty. More recently, an improved stochastic programming called robust programming has been developed with the capability of tackling this shortage. Owing to the flexible modeling qualifications allowed by robust optimization (RO), it is believed that this approach can provide a credible methodology for real-world uncertain logistics problems. In other words, the simplicity of implementation of this method enables decision makers to manage and control the logistics system without having to learn complicated programming procedures.