Abstract: Agent-based modelling of human crowds has now become an important and active research eld, with a wide range of applications such as military training, evacuation analysis and digital game. One of the signicant and challenging tasks in agent-based crowd modelling is the design of decision rules for agents, so as to reproduce desired emergent phenomena behaviors. The common approach in agent-based crowd modelling is to design decision rules empirically based on model developer’s experiences and domain specic knowledge. In this project, genetic programming based approach is proposed to automatically learn decision rules for agent-based crowd models, so as to reproduce desired crowd behavior. Simulation results have demonstrated the feasibility of the approach and shows that our algorithm is able to find decision rules for agents, which in turn can generate the prescribed macro-scale dynamics.
Related paper:
- J. Zhong, L. Luo, W. Cai, and M. Lees, “Automatic Rule Identification for Agent-Based Crowd Models Through Gene Expression Programming,” In Proceedings of the 2014 International Conference on Autonomous Agents and Multi-agent Systems (AAMAS 2014), pp.1125-1132, International Foundation for Autonomous Agents and Multiagent Systems, 2014.
- J. Zhong and Wentong Cai, “A Hyper-Heuristic Framework for Agent-Based Crowd Modeling and Simulation,” 2016 International Conference on Autonomous Agents and Multiagent Systems (AAMAS2016), pp.1331-1332.
- J. Zhong, W. Cai, M. Lees, L. Luo, “Automatic Model Construction for the Behaviour of Human Crowds,” Applied Soft Computing, Vol.56, pp.368-378, July, 2017.
Related videos for submitted manuscript:
D. Li and J. Zhong, “Dimensionally Aware Multi-objective Genetic Programming for Automatic Crowd Behavior Modeling”, manuscript, submitted
Video1: Static-DC
Video2: Combined-DC