Design and Prototyping An Agent-based Intelligent Disaster Management System Based on Collaboration of Heterogeneous Information Systems

Abstract

Natural disasterssuch as earthquakes, tsunamis, floods, storms and volcanic eruptions, cause severe damage to millions of people and their living infrastructures; they also have a devastatingimpact on the country’s economic and political sectors. If these disaster are not managed properly they would lead to political and social instability, they would also affect the security and international relations too. The response phase is the most complex phase of disaster management. This phase is usually accompanied with high stress, environment change tolerance and lack of information. Since thebeginning of the disaster, disaster assessment process is initiated. And delays in accessing the right information and wrong decision making, is considered as one of the factors leading to the failure of the rescue groups. This same factor usually leads to the intensification of consequencesand damages. Subsequently, in the last two decades the development of disaster management systems hasgained the attention of researchers and governments such as the United States, Europe andChina. And community-based joint information system, ontology-enriched multi-documentsummarization, and fuzzy logic based decision support System are considered among these kindof researches. But then again past researches lack total solution for the assessment and decisionmaking. Another important point is that an information system (unintelligent software) is sometimes faced with a disaster management environment, which in itself is an environment without accessibility, uncertain, and dynamic. This environment may sometimes cause the system to stop and bring about irreparable consequences.Due to the disaster environment conditions there is a need for the use of intelligent agents instead of an information system. So that in case of drastic and unpredictable changes and in the open environments, where it is likely that an activity lead to failure, they work together as a multiagent and non-stop system. The proposed model and system in this study included a group of intelligent agents that can be used alongside disaster management team for disaster assessment and providing solutions to exit the disaster. This system has unique characteristics such as independence of knowledge from code, customization, scalability, communication with similar versions, distributedstructure and fault-tolerant. Although this system is designed for disaster management, it has the capability to be trained and used in various management fields