Recently, fire evacuation problems in high-rise buildings have been increasing due to increased occupant density and limited evacuation routes, leading to increased casualties. To address this issue, this study proposes a distributed evacuation performance evaluation method that considers floor-by-floor distribution ratios. Using FDS+EVAC-based simulations, various floor-by-floor distribution scenarios were designed and their impact on evacuation times and bottlenecks was quantitatively analyzed. The results confirmed that rationally adjusting the floor-by-floor exit and elevator utilization ratios shortened overall evacuation times and reduced congestion in specific areas. This method can serve as a foundation for establishing evacuation plans and fire safety design standards for high-rise buildings.

A*algorithm is highly useful to search the shortest route to the destination in the evacuation simulation. For this reason, A*algorithm is used to evaluate the evacuation experiment by the computer simulation. However there are some problems to analyze the outcome in relation to the reality. Because all the people in the building are not well-informed of the shortest route to the exit. And they will not move to the disaster spot though it is shortest route to the exit. Therefore, evacuation simulation program based on A*algorithm raise a problem of bottleneck phenomenon and dangerous result by damage surrounding the disaster spot. The purpose of this research is to prove the necessity for dispersion evacuation simulation by Multi agent system to solve the problems of the existing evacuation simulation program based on A*algorithm.