A multiagent architecture for a distributed artificial intelligence application

Abstract

Artificial intelligence applications are normally complex, demanding much computing power. This paper presents the multiagent architecture of StormCast, a distributed artificial intelligence application for severe storm forecasting. Important objectives as scaleability, robustness and reusability can be met by partitioning an ove rall task in subtasks, distribution of the subtasks on different nodes, parallel processing of the subtasks and cooperation between the subtasks.