Distributed and guided genetic algorithm for humanitarian relief planning in disaster case

Abstract

In this paper we propose a distributed and guided genetic algorithm for humanitarian relief planning in natural disaster case. It is a dynamic vehicle routing problem with time windows (DVRPTW), where customers should be served during a given time interval. This problem is an extension of classic vehicle routing problem. In the case of a disaster, emergency planning must be fast, consistent and scalable. For these reasons we opted for an improved genetic algorithm by adding some sort of guide to accelerate the convergence of the algorithm. Thus, the genetic algorithm can provide a population of solutions that can address the dynamic aspect of the problem. The objective of our approach is to provide a plan to meet all the demands with minimizing the total distance travelled. The proposed approach has been tested with theoretical data and showed high efficiency, which infers the possibility of applying for the management of emergency calls in the event of major disaster. © Springer International Publishing Switzerland 2014.