Multi-Period Vehicle Routing & Replenishment Problem of Neighbourhood Disaster Stations for Pre-Disaster Humanitarian Relief Logistics

摘要

Natural disasters are uncontrollable situations that affect human life directly. Despite the researches and technological progress, it is still not possible to predict when or where the natural disaster will occur beforehand. Natural disasters cause severe loss of lives and damages. In addition, they cause physical, financial, social and environmental losses. Pre-disaster, during disaster and post-disaster activities are significant in order to decrease the losses caused by natural disasters. This study is about one of the pre-disaster activities. In the pre-disaster management process, a new activity is being tried in Turkey. Called "Neighborhood Disaster Stations", containers filled with emergency relief items such as medicines, painkillers, antiseptics and canned goods are located at the different predetermined locations. It is important to keep items in these stations usable at all times. Since these commodities have expiration dates, they need to be replenished periodically in order to remain useable at any time. The proper time for the replenishment should be determined by considering the probability of reselling or re-using the commodities with the maximum return as much as possible. However, replenishing frequently will result in large operational costs. Therefore, there is a trade-off between routing costs and replenishment. We propose a novel mixed integer-programming model in order to solve this problem. The proposed model determines the replenishment policy for each commodity in the containers and generates the route of each vehicle within a given planning horizon. The objective of this study is to maximize the total profit, which is the difference between expected revenue from reselling and the transportation cost for total routing costs for time periods in the planning horizon. The model determines the replenishment date of each commodity in each disaster container and provides optimal route for each vehicle within planning horizon. The proposed mixed integer programming model is solved optimally for a small instance in IBM ILOG CPLEX Optimization Studio 12.8 and validation of the model is done.