A Fuzzy Logic Inspired Cellular Automata Based Model for Simulating Crowd Evacuation Processes

This work investigates the incorporation of fuzzy logic principles in a cellular automata (CA) based model that simulates crowd dynamics and crowd evacuation processes with the usage of a Mamdani type fuzzy inference system. Major attributes of the model that affect its response, such as orientation, have been deployed as linguistic variables whose values are words rather than numbers. Thus, a basic concept of fuzzy logic is realised. Moreover, fuzzy if-then rules constitute the mechanism that deals with fuzzy consequents and fuzzy antecedents. The proposed model also maintains its CA prominent features, thus exploiting parallel activation of transition rules for all cells and efficient use of computational resources. In case of evacuation, the selection of the appropriate path is primarily addressed using the criterion of distance. To further speed up the execution of the Fuzzy CA model the concept of the inherent parallelization was considered through the GPU programming principles. Finally, validation process of the proposed model incorporates comparison of the corresponding fundamental diagram with those from the literature for a building that has been selected for hosting the museum ‘CONSTANTIN XENAKIS’, in Serres, Greece.