A multi-agent-system framework for flooding events

This paper presents the potential capabilities offered by an integrated multi-agent system comprising logical agents and a neural network, specialized in monitoring flood events for civil protection purposes. Here we describe the idea of a framework – at the moment only partially developed – consisting of a set of intelligent agents, which perform various tasks and communicate with each other to efficiently generate alerts during flood crisis events, collaborating with a neural network, derived from the PSP-Net model, which is dedicated to the inspection and analysis of satellite images.