2014/12/16 by Sankha Baran Dutta, Dutta, Sankha Baran, R.D. McLeod +4 · 1 citation
Engineering · Social Sciences · #Evacuation and Crowd Dynamics #Human Mobility and Location-Based Analysis #Transportation Planning and Optimization
paper · pdf · doi:10.48550/arxiv.1412.4933
Pedestrian movement, although ubiquitous and well-studied, is still not that\nwell understood due to the complicating nature of the embedded social dynamics.\nInterest among researchers in simulating pedestrian movement and interactions\nhas grown significantly in part due to increased computational and\nvisualization capabilities afforded by high power computing. Different\napproaches have been adopted to simulate pedestrian movement under various\ncircumstances and interactions. In the present work, bi-directional crowd\nmovement is simulated where an equal numbers of individuals try to reach the\nopposite sides of an environment. Two movement methods are considered. First a\nLeast Effort Model (LEM) is investigated where agents try to take an optimal\npath with as minimal changes from their intended path as possible. Following\nthis, a modified form of Ant Colony Optimization (ACO) is proposed, where\nindividuals are guided by a goal of reaching the other side in a least effort\nmode as well as a pheromone trail left by predecessors. The basic idea is to\nincrease agent interaction, thereby more closely reflecting a real world\nscenario. The methodology utilizes Graphics Processing Units (GPUs) for general\npurpose computing using the CUDA platform. Because of the inherent parallel\nproperties associated with pedestrian movement such as proximate interactions\nof individuals on a 2D grid, GPUs are well suited. The main feature of the\nimplementation undertaken here is that the parallelism is data driven. The data\ndriven implementation leads to a speedup up to 18x compared to its sequential\ncounterpart running on a single threaded CPU. The numbers of pedestrians\nconsidered in the model ranged from 2K to 100K representing numbers typical of\nmass gathering events. A detailed discussion addresses implementation\nchallenges faced and averted.\n