You are here

Michalis Fragkos

2017 Diploma Thesis Title:Supply Planning in Natural Disasters:Modelling and Analysis                                                                                                                                            


The thesis deals with modeling and analysis of supply planning during or immediately after a natural  disaster.  In  post  emergency  response  planning,  the  supply  of  consumable  and  nonconsumable  provisions  for  both  civilians,  who  evacuate  residential  areas,  and  intervention groups  at  the  corresponding  shelters,  is  of  immediate  importance.  In  this  thesis,  provisions supply is modeled and analyzed by introducing  the  Emergency Supply using Heterogeneous Fleet Problem (ESHFP).
Initially, a Mixed Integer Linear Programming (MILP) mathematical model is introduced for the ESHFP. In order to solve this problem, we have developed a  novel heuristic algorithm, which aims in determining the set of routes and the vehicles that can be used to minimize the total supply time, respecting constraints concerning routing, timing, capacity and supply. 
Since the corresponding MILP is difficult to be solved to optimality in reasonable time, we have introduced a novel heuristic approach for ESHFP which minimizes the total time needed to collect provisions from available pick up locations  and (by using appropriate vehicles  among those available)  to deliver provisions to  a)  evacuees at shelters and  b)  intervention groups  at their accommodation sites.  The proposed heuristic approach takes into account all necessary constraints described in the MILP model. 
To validate the effectiveness this approach, we have applied the proposed algorithm to a series of examples, generated randomly.  Furthermore, we have used the proposed algorithm to deal with a real case study involving a significant forest fire in the Province of Teruel in Spain. The results  of  both  the  tests  and  the  case  study  are  very  encouraging,  attesting  to  the comprehensiveness of the proposed model and the efficiency of the new solution heuristic.