Athens University Economics and Business (AUEB)
Wildfire simulation software for intermittently connected networks
Abstract
dc:descriptionThis thesis presents three contributions to the development of modular wildfire simulation software. The first is a Wildfire Simulation Ontology-WSO, a domain ontology implemented as a software module that provides structured representations of the environmental components relevant to wildfire simulation. WSO addresses the heterogeneity inherent in crowdsourced observations arriving in diverse geometric formats, at varying spatial resolutions, and with inconsistent attribute vocabularies. Its aim is to normalize them into semantically consistent, GeoJSON encoded outputs that can be directly consumed by downstream simulation software components. The second component is a data fusion pipeline that resolves the record linkage problem arising when multiple participants independently describe the same feature at overlapping or proximate times. Raw crowdsourced data are inherently redundant and geometrically inconsistent, so they must be linked to a single representative geometric representation before they can serve as simulation inputs. We address this problem as a two-stage density-based clustering of WSO instances, followed by a contour correspondence process based on Ant Colony Optimization that reconstructs a semantically consistent representation of the geometric primitives that can serve as an input to the simulation engine. The third software component is a real-time calibration engine that keeps the simulator continuously aligned with observed fire behavior as new field data arrives. The wildfire simulator is implemented as a Cellular Automata (CA) model, in which the temporal evolution of simulation cells is governed by a set of continuous parameters encoding fuel ignitability, wind coupling coefficients, and topography. In conventional software, these parameters are fixed and cannot be adapted to reduce any divergence between the simulated and observed fire behavior. We propose a dynamic fine-tuning mechanism based on Ant Colony Optimization for Continuous Domains that searches the parameter space to identify the optimal parameter configuration, and show that it can be applied to diverse CA-based simulation models.
Degree
thesis:*- Grantor dc:publisher
- Athens University Economics and Business (AUEB)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Karakonstantis, Ioannis
- Καρακωνσταντής, Ιωάννης
Subjects
dc:subject × 18- Προσομοίωση πυρκαγιάς
- Κυτταρικά αυτόματα
- Βελτιστοποίηση αποικίας μυρμηγκιών
- Πληθοπορισμός
- Δίκτυα διαλείπουσας συνδεσιμότητας
- Wildfire simulation
- Cellular automata
- Ant colony optimization
- Crowdsourcing
- Intermittently connected networks
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Λογισμικό (software)
- Δίκτυα ηλεκτρονικών υπολογιστών και Επικοινωνίες
- Computer and Information Sciences
- Natural Sciences
- Software
- Computer Networks and Communications
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62827