مدل های شبیه سازی تخلیه: چالش ها در مدل سازی افزایش تخلیه ساختمان با رویکرد اتوماسیون سلولی

Building evacuation simulation provides designers with an efficient way of testing the safety of a building before construction. A significant number of models have been developed in a variety of disciplines (computer graphics, robotics, evacuation dynamics, etc.). This paper presents a review of crowd simulation models and selected commercial software tools for high rise building evacuation simulation. The commercial tools selected (STEPS and EXODUS) are grid-based simulations, which allow for efficient implementation but introduce artifacts in the final results. This paper focuses on describing the main challenges and limitation of these tools, in addition to explaining the importance of incorporating human psychological and physiological factors into the models. The paper concludes with an overview of fundamentals that should be applied to simulate human movement closer to real movements of people, where interaction between bodies emerges and flow rates, densities, and speeds become the result of those interactions instead of some predefined value.

There have been many models developed in order to provide designers with ways of forecasting evacuation times for buildings. A large number of models for pedestrian simulation have been developed over the years in a variety of disciplines (computer graphics, robotics, evacuation dynamics, etc.). These models can be classified into two subsets: macroscopic and microscopic models. Macroscopic models focus on the systems as a whole while microscopic models study the behavior and decisions of individual pedestrians and their interaction with other pedestrians in the crowds. Macroscopic models include regression models, route choice model, queuing models, and gas-kinetics models. Regression models use statistically established relations between flow variables to predict pedestrian flow operations under specific circumstances. The characteristics of this flow depend on the infrastructure (stairs, corridors, etc.) [15]. Route choice models describe pedestrian wayfinding based on the concept of utility. Pedestrians choose their destinations in order to maximize the utility of their trip (in terms of comfort, travel time, etc.) [11]. Queuing models use Markov-chain models to describe how pedestrians move from one node of the network to another. Nodes are usually rooms, and therefore links are usually portals or doors. Markov-chain models are defined by a set of states together with transition probabilities. At each extrapolation step, a successor state is selected by either sampling from the transition distribution, or identifying the most probable successor [12].

Even though current software based on cellular automata approaches has been widely used and validated, there is still a need to develop models that can closely simulate human behavior. It is important to consider physical interactions between individuals and the resulting impact of these interactions in the behavior of the virtual humans (trajectories, bottlenecks, flow rates, etc.). On top of this improved human movement model, it is essential to integrate agent-based approaches that endow the agents with physiological and psychological elements that can describe the virtual humans' mental state at any given time and drive their decision making, orientation skills and possible roles within the crowd. Communication between agents is another crucial element to integrate in realistic models to achieve more accurate evacuation results. From the data gathered in the literature, it can be observed how communication has an important impact on the evacuation since it actually drives when the evacuation starts. Crowd simulation systems are bridging the gap between real evacuation performance and the tedious process of calculating the evacuation times following engineering guides for safe egress. At present, these tools provide the designer with graphical information of where the bottlenecks could appear, and also make easier the study of how some changes in the building could modify the overall evacuation. Even though these systems have been validated for some particular scenarios, further work needs to be carried out to improve the accuracy of the results.