deletions | additions       diff --git a/untitled.tex b/untitled.tex  index 98a9b82..b4ebf1b 100644  --- a/untitled.tex  +++ b/untitled.tex 
...
\section{A multi-objective relief chain location distribution model for urban disaster management}  Barzinpour and Esmaeili \cite{BarzinpourEsmaeili2014} study the mitigation of consequences of natural disasters on affected people. They argue the importance of location/allocation models in humanitarian supply chains that aim to provide quick relief items in the right time and the right place. In particular, they suggest a multi-objective mixed-integer linear program (MILP) solved on a grid representation of the city (using geographic information system (GIS)), which determines the location of distribution centers (DCs) and the assignment of DCs to the demand points. The authors suppose that the demand of each grid is a function of its population and its potential damage severity. To provide the necessary input for their model, they have employed a special-purposed software that assess the seismic vulnerability in cities. In their model, the authors consider various type of costs such as setup costs, holding costs, transportation costs, and shortage costs. They also consider three objective functions: 1) maximization of the cumulative coverage of population, 2) minimization of the total setup costs, and 3) minimization of the total transportation, holding, and shortage costs. To solve the problem, the authors have used the goal-programming approach. In particular, they have used expert judgments for the ideal values of the objective function. Furthermore to be able to tackle problem, the authors have assumed that the district can be divided into ten subregions, where the DC in each subregion will only serve within the subregion. This results in a significant reduction into the model size and allows the usage of standard optimization solvers. In particular, they report the result of their model for the district with 350,000 inhabitants, 10 subregions, and four types of relief products.   \section{A robust optimization approach to post-disaster relief logistics planning under uncertainties}  Liu and jiang \cite{LiuJiang2015} mentioned the importance of emergency logistics plan.   A post-disaster study is conducted by the authors on the delivering relief personnel (i.e., health personnel and aid workers) and evacuating critical population in mountainous areas. By considering uncertainty in affected area's demand of relief and critical population and also transportation time between temporary facilities (TFs) and affected areas (AAs), they proposed a bi-objective robust optimization model, which determines the level of relief supplies at TFs and overall transportation plan between TFs and AAs. Due to mountainous nature of the region, the authors assume that TFs set up at the rim of the affected area and all the transportation will be done by helicopters. In their model, the authors comprised total weighted unsatisfied demand in the first objective and the sum of the total costs of mobilization of relief supplies, helicopter recruiting and transportation is included in the second objective. The authors handled lexicographic optimization technique to solve the problem, in which two sequential single objective model, with priority of the first objective, is solved. For evaluating their model they conducted a case study on data extracted from Great Sichuan earthquake.   \section{A two-echelon stochastic facility location model for humanitarian relief logistics}  Doyen et al. \cite{DoyenEtal2011} elaborated necessity of pre- and post-disaster planning in relief logistics of earthquake disaster.  They developed a two-echelon stochastic facility location model for pre-positioning relief(or essential) items and proper distribution of these items after an earthquake. In the model pre- and post- disaster decisions constituted the first and second stage of the model, respectively. In the first stage location of the regional rescue centers (RRCs) and the amount of items stocked in each RRC is determined, while in the second stage decisions about locations of the local rescue centers(LRCs), LRC-demand point assignment, flow amount at both of the echelons and shortage amount are made. Therefor, the relief items firstly stored in RRCs in pre-disaster phase and then are distributed to the demand points through LRCs that are set up in the post-disaster stage. The post disaster stage function as the recourse variable as it incorporated uncertainty by defining scenarios on possible earthquake intensity. In the model, minimization of total facility location, inventory holding at RRCs, transportation at each echelon and demand shortage costs is intended. The authors assumed that RRCs are uncapacitated and the LRCs are capacitated, each LRC/demand point can receive items from a single RRC/LRC. In addition, in the model, demand shortage is captured by imposing high penalty cost. The solution procedure utilized Lagrangian relaxation heuristic (LH) armed with local search algorithm, in which LH applied by relaxing joining constraints of the two echelons to obtain a lower bound and an upper bound is calculated by fixing location variables, then local search algorithm iteratively performed in the neighboring of the solution to further improve it.   The performance evaluation is carried out on a problems ranging from {5,10} RRCs, {10,20,40}, {50, 75, 100, 200, 400, 600, 800} demand points and two scenario types each of with four scenarios. Computational comparison also is conducted with commercial solvers in term of gap reduction and value of stochastic solution (VSS).   the experiments shows that the procedure significantly outperforms commercial solvers.