Application of a multiagent system for resource distribution in humanitarian logistics

Abstract

In humankind’s history, many natural and man-made disasters have happened and will continue to happen. In a post-disaster scenario, many processes and actors must be organized to provide the needed supplies to the affected people. Although it is similar to a supply chain, it is called Humanitarian Logistics in a disaster context. Many papers present centralized approaches to organizing different aspects of humanitarian logistics. However, a centralized approach is not always appropriate due to the dynamics and uncertainty present in a post-disaster scenario. This paper describes a distributed approach based on Multiagent Systems (MAS) to organize the elements of humanitarian logistics in a post-disaster scenario (MAS-HL). Intelligent agents in the MAS-HL represent real-world elements, such as affected zones, distribution centers, donors, and trucks. The agents negotiate among themselves using a Contract-Net Protocol (CNP) to organize the operations of the elements represented. Due to dynamics and uncertainty in a post-disaster scenario, the agents must reorganize the operations of the elements. The MAS-HL is evaluated by simulating and evaluating a case study focused on bottled water delivery in post-disaster scenarios, they represent when major changes occur. The computational results confirmed that the MAS-HL quickly generates plans to organize the elements of humanitarian logistics and reorganize the plans when the environment changes.