Ndiaye, MalickAbougharib, Arwa2020-08-252020-08-252020-0735.232-2020.27http://hdl.handle.net/11073/19726A Master of Science thesis in Engineering Systems Management by Arwa Moamen Abdelfattah Abougharib entitled, “Design of A Theoretical Framework for A Real-Time Fire Evacuation Guidance System”, submitted in July 2020. Thesis advisors is Malick Ndiaye. Soft copy is available (Thesis, Approval Signatures, Completion Certificate, and AUS Archives Consent Form).Despite modern-day advancements in fire protection technologies, stricter fire codes, and more sophisticated modeling of fire propagation and human behavior, fire alarm systems are notably losing the ability to induce evacuative behavior, as people have habituated and grown unresponsive to false alarms. Moreover, people still require active support and guidance during evacuation from a building under fire and are still at risk of using suboptimal routes that could open to untenable conditions. The risk is higher for residential buildings where the majority of all civilian deaths due to fire occur. Therefore, this thesis proposes a smart notification and guidance system with an IoT architecture that can significantly reduce pre-movement and movement times by supporting the decision-making process of evacuees during a fire disaster in a residential building. At the front-end, the system uses crowd, temperature, and smoke sensors to measure environmental conditions, and communicates with occupants through a mobile application. At the back-end, a rule-based optimization approach is used to optimally route evacuees while avoiding the dangers of the fire. The system must also test the recommended egress paths for future tenability by comparing estimated travel times against the Available Safe Egress Time computed by a thermal fire simulator. This thesis develops the theoretical framework and lays the cornerstones for such a system concept through six algorithms, three of which are for pre-optimization processing. The building network problem afterward is framed as an earliest arrival problem and solved using a strongly polynomial earliest arrival transshipment algorithm. As by-products of this work, extensions to the hydraulic model are proposed to account for the effects of smoke and crawling behavior during egress using empirical data.en-USFire evacuationFire protectionFire alarm notificationHuman behavior in firesBuilding network modelingEarliest arrival evacuationHydraulic modelThesis