Abstract:

Establishing a two-layer optimization framework based on genetic algorithm and pedestrian simulation model, a single-chamber dual-exit scene is selected, in which two exits located on the same side, on the adjacent sides, and on the opposite sides are set as three typical layouts. Then three types of body size, three conditions of initial crowd density, and nine kinds of evacuation exits’ widths are used to construct 243 numerical experimental scenarios, and the optimized exit layouts which are most beneficial for pedestrian evacuation are obtained respectively. It finds that the evacuation exit width is positively related to the evacuation efficiency, and the pedestrian radius as well as initial crowd density is negatively related to the evacuation efficiency. Moreover, the key parameters that affect the evacuation efficiency are different in varied layouts.

Key words: pedestrian evacuation, layout optimization, social force model, genetic algorithm

摘要：

构建基于遗传算法与行人运动仿真模型的双层优化框架，选取单室双出口场景，考虑两个出口在同一侧、相邻侧、相对侧3种典型布局，设定3种人员身形，3种疏散初始人群密度，9种疏散出口大小条件，形成243个数值实验场景，分别计算获得最有利于人员疏散的出口布局。研究发现疏散出口宽度与疏散效率呈正向关系，疏散人员半径、初始人员密度与疏散效率呈负向关系，同时在不同布局中，影响疏散效率的关键参数有一定区别。

关键词: 行人疏散, 布局优化, 社会力模型, 遗传算法