Back to search

University of Thessaly (UTH)

Agent-based coordination for first responders in disaster management leveraging AI and semantic web technologies

Abstract

dc:description

Natural disasters impose significant human, economic, and environmental burdens, disrupting communities and overwhelming critical infrastructure. In recent years, their growing frequency and intensity have placed unprecedented demands on disaster risk management (DRM), requiring rapid coordination among multiple stakeholders and effective allocation of limited resources. Within this context, the response phase of DRM is especially critical, as decisions must be made under time pressure, incomplete information, and high uncertainty. Information and communication technologies (ICT) have played an important role in this direction by improving the collection and exchange of disaster-related information. However, current systems continue to face difficulties in integrating heterogeneous data into a consistent semantic representation. This limitation reduces their ability to interpret data, infer new knowledge, or adapt past cases to current operational needs. To overcome this challenge, systems should combine semantic integration with adaptive reasoning to deliver recommendations responsive to evolving disaster conditions. This thesis introduces the First Responders System (FiReS), a framework that combines Semantic Web technologies with agent-driven Case-Based Reasoning (CBR) to support decision-making in post-disaster scenarios. FiReS is organized into two layers. The Design Layer establishes the semantic foundation through the FiReS ontology, which defines disaster-related entities, integrates heterogeneous data, and applies SWRL rules for reasoning. The Execution Layer builds on this foundation to generate recommendations. The FiReS agent manages the CBR cycle by retrieving past cases from the RDF repository, comparing them through multi-dimensional similarity assessment, and applying Bayesian reasoning to address uncertainty. Adapted cases are validated through SWRL-based reasoning to ensure logical consistency and then stored back into the RDF repository, supporting iterative supporting iterative learning within the system. FiReS was validated in two stages. The Design Layer was evaluated as the semantic foundation of the system. The FiReS ontology provided a consistent representation of heterogeneous disaster data and was applied in predictive modeling, classification, and time-series forecasting. In addition, the use of SWRL rules confirmed that domain-specific logic could be applied reliably within the semantic framework. The Execution Layer was validated to examine whether it could retrieve, adapt, and refine past cases aligned with the needs of new disaster scenarios. Its performance was measured with two metrics: efficiency scores and matching rates. The results showed that the Execution layer maintained high efficiency while also covering a broad range of attributes, indicating that the FiReS agent can provide recommendations that remain both reliable and adaptable under varying conditions. In conclusion, this thesis demonstrates that effective disaster response requires systems capable of adaptive reasoning under uncertainty and time-critical constraints. FiReS addresses this challenge by integrating Semantic Web technologies with agent-based reasoning into a unified framework that delivers semantically consistent recommendations that align with the operational needs. Future research will focus on testing the system in real-time disaster operations and explore how it can be extended to support other phases of disaster management.

Degree

thesis:*
Grantor dc:publisher
University of Thessaly (UTH)
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Bania, Areti
  • Μπανιά, Αρετή

Subjects

dc:subject × 24

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/62138

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Bania, Areti; Μπανιά, Αρετή. Agent-based coordination for first responders in disaster management leveraging AI and semantic web technologies. University of Thessaly (UTH), 2026. http://hdl.handle.net/10442/hedi/62138