熵修正的混合人工蜂群-蝙蝠算法人群疏散模型

目前的群智能疏散模型多仅考虑单一的经典的群体智能,不足以描述复杂的群体疏散行为特征,且鲜有考虑人群混乱程度对人群疏散的影响。为研究描述多种群体疏散行为的群智能疏散模型,综合使用多种群智能算法,并考虑了人群混乱程度对疏散的影响,构建了熵修正的混合人工蜂群-蝙蝠算法人群疏散模型。首先,采用DBSCAN（density-based spatial clustering of applications with noise）算法进行群组划分。然后,将人群分为群组引导者、群组成员和离散人员3类,并针对每类人群的特点,基于蝙蝠算法描述群组引导者,基于人工蜂群算法描述群组成员,基于粒子群算法描述离散人员。最后,引入定量描述人群混乱程度的疏散熵对群组引导者进行位置修正,构建了熵修正的混合人工蜂群-蝙蝠算法人群疏散模型。仿真结果表明,该模型可以模拟群组疏散,比较符合真实的群组疏散形状,以群组形式疏散一定程度提高了疏散效率；同时,引入疏散熵进行修正后,群组引导者可以引导群组成员避开前方混乱区域,避免了人群过度集中,增强了疏散的安全性与快速性。

Hybrid artificial bee colony-bat algorithm-based evacuation model with entropy correction

Current swarm intelligence evacuation models only consider single classic swarm intelligence, which is insufficient to describe the complex behavior characteristics of crowd evacuation. In addition, these models rarely take into consideration the impact of crowd chaos on crowd evacuation. In order to study the swarm intelligence evacuation model describing the evacuation behaviors of different groups, by integrating various swarm intelligence algorithms, and taking into account the impact of crowd chaos on evacuation, a crowd evacuation model based on hybrid artificial bee colony-bat algorithm with entropy correction was proposed. Firstly, the density-based spatial clustering of applications with noise (DBSCAN) algorithm was used for group partition. The evacuees were divided into group leader, group members, and disorganized people. Next, according to the characteristics of each type of evacuees, the group leader was described based on bat algorithm, the group members based on artificial bee colony algorithm, and the disorganized people based on particle swarm optimization (PSO). Finally, the evacuation entropy that quantitatively describes the degree of crowd chaos was introduced to correct the position of the group leader, and the evacuation model based on hybrid artificial bee colony-bat algorithm with entropy correction was thus constructed. Simulation results show that the model could well simulate group evacuation, which was basically consistent with the real shape of group evacuation, and the evacuation efficiency was improved by means of group evacuation to some extent. With the introduction of the evacuation entropy for correction, the group leader could guide the group members to avoid the chaotic area ahead, prevent the excessive concentration of evacuees, and enhance the safety and rapidity of evacuation.