Investigation of pedestrian evacuation by using ant algorithm

It is observed that pedestrians in an emergency exhibit collective behavior. In this work, this collective behavior was modeled by using ant algorithm. The ants in a colony communicate with each other by using a chemical, called pheromone. They dropped pheromone on the substrate crawling forward so that they can find the shortest route between the nest and the food. It is seen experimentally that ants, in general, follow the route involving more concentration of pheromone. With this point of view, in this work, each person dropped a 'visual pheromone' on the floor and after a while, the most preferred routes constituted an 'attractive force' for the other people. As the pheromone that is used by ants evaporates, these 'virtual pheromone' evaporates with time too. The evacuation process was analyzed for a room. The evacuation process was investigated for three different situations: 1. Any emergency where there is neither smoke and nor fire; 2. In the event of a fire that has occurred for any reason, a situation in which only smoke is present in the room and there is limited contact due to smoke; 3. Both fire and smoke exist at a time. In evacuation process, one of the possible reasons to reduce the communication among people is smoke. In this study, it was assumed that the amount of smoke was not distributed inhomogeneously in the field and its flow was not laminar for the situations 1. and 2. The propagation of smoke in the room was simulated by using sandpile model. The Monte Carlo simulations for the 3rd situation were performed under consideration that a) fire did not propagate, b) fire propagated. The evacuation time was investigated for certain values of parameters (evaporation rate of pheromones, the location of doors, obstacles and fire, the movement of smoke) used in the proposed model by performing Cellular Automata Monte Carlo computer simulations.