CROS: A Contingency Response multi-agent system for Oil Spills situations

Abstract

This paper presents CROS, a Contingency Response multi-agent system for Oil Spill situations. The system uses the Case-Based Reasoning methodology to generate predictions to determine the probability of finding oil slicks in certain areas of the ocean. CBR uses past information to generate new solutions to the current problem. The system employs a SOA-based multi-agent architecture so that the main components of the system can be remotely accessed. Therefore, all functionalities (applications and services) can communicate in a distributed way, even from mobile devices. The core of the system is a group of deliberative agents acting as controllers and administrators for all applications and services. CROS manages information such as sea salinity, sea temperature, wind speed, ocean currents and atmosphere pressure, obtained from several sources, including satellite images. The system has been trained using historical data obtained after the Prestige accident on the Galician west coast of Spain. Results have demonstrated that the system can accurately predict the presence of oil slicks in determined zones after an oil spill. The use of a distributed multi-agent architecture has been shown to enhance the overall performance of the system.