Abstract: In order to minimize the casualties and losses caused by fire in the exhibition hall. This paper puts forward research on fire evacuation path of exhibition hall based on improved ant colony algorithm. Firstly, quantified fire products, angle and distance factors are used to improve the heuristic function. The pheromone update model conforming to the process of crowd evacuation was established. Thereby, the traditional ant clony algorithm was improved. Secondly, the dynamic evacuation model of personnel was established, with the shortest evacuation time and distance as the target function; the two-dimensional evacuation environment and the grid map are built, and then rises to the 3D environment model. Finally, taking an exhibition hall in Xi'an as an example to conduct research, the 3D grid environment of 30×25×3 was established in MATLAB, the parameters were optimally combined and configured. The improved ant colony algorithm was used to simulate fire scenes at different stages and the convergence was analyzed. Results show that the improved ant colony algorithm is suitable for 3D evacuation models. The algorithm has excellent convergence speed and search capabilities. Using this improved algorithm in large exhibition hall, the best evacuation route can be quickly found. It helps to improve the efficiency of personnel evacuation in exhibition hall.

Key words: improved ant colony algorithm, 3D evacuation environment model, dynamic simulation of the fire, personnel evacuation efficiency