deletions | additions       diff --git a/section_A_three_stage_stochastic__.tex b/section_A_three_stage_stochastic__.tex  index c2b9019..351b1a0 100644  --- a/section_A_three_stage_stochastic__.tex  +++ b/section_A_three_stage_stochastic__.tex 
...
\section{A three-stage stochastic facility routing model for disaster response planning}  Rennemo et al. 2014 \cite{RennemoEtal2014} consider the problem of locating and routing in humanitarian logistic and present a three-stage stochastic model for planning and distribution of relief items. To capture real-world setting, demand of affected area, vehicle fleet capacity and state of the transportation network are considered as uncertain parameters and are handled by several discrete scenarios.  The proposed model include three-echelon distribution centers: 1) International Central Depots (ICDs)), 2) the Local Distribution Centers (LDCs) and 3) the different Points Of Distribution (PODs). ICDs are initial suppliers, LDCs that act as drop points and intermediate suppliers and POSs are centralized locations for delivering relief items. The model is then formulated as three-stage mixed integer multi-commodity multi-vehicle stochastic problem; once a disaster happen, first stage decide to-be-open LDCs, allocation of commodities to LDCs and number of vehicles arrived at LDCs. Already made decisions in former stage as well as information about demands at PODs, make transition from first stage to second. Thus, at the beginning of second stage, vehicle fleet capacity of each LDC as well as demands at AAs, that were uncertain at the onset of disaster, are deterministic but condition of transportation network is still uncertain. At this point, vehicles are loaded and dispatched from the LDCs to PODs through generated initial routes. Since the condition of road is unknown yet, third stage is responsible for taking recourse action for already sent vehicles; if a road preserves its functionality the delivery is made according to second stage decision. Otherwise, a vehicle encounters an obstacle and a modified route is generated for the vehicle.   In their computational study,