Human Crowd Simulation: What Can We Learn From ALife?

One of the key components of the suspension of disbelief in real-time 3D simulations is the apparent authenticity of actions and gestures played by the individuals of the virtual population. This paper addresses this aspect of simulation, by investigating ways to improve the behavioral realism of virtual humanoid characters in groups and small multitudes. We look at the framework of ALife, identifying and analyzing existing bio-mimicking techniques that can be used in this context and contribute towards the improvement of the plausibility from the generated simulations. By looking at the literature, we identify some of the key elements from ALife that are being progressively incorporated in the simulations of groups and crowds. Then, we discuss a generative model for spontaneity and heterogeneity where bio-inspired agents are individualized with DNA-like strings and appear organized hierarchically exchanging token units of energy, mass, and resources. The result is a generative population of agents that self-organize and interact autonomously, exhibiting interesting social dynamics based on biological tenets and an economy of resources. We analyze this simulation quantitatively with the purpose of studying the impact of each of the previously identified techniques.