Studying spatial effects on human mobility patterns using agent-based simulations

1. Random and goal oriented walk models To simulate human movement in a street network, we setup some random walkers which can hop from one street to another arbitrarily. The hopping behaviour is defined at a topological level in which individual streets are represented as nodes and street intersections as links of a connectivity graph (Figure 1). On the other hand, the simulation has yet to be based on a geometric level in order to mimic a sort of network constrained movement. That is, within an individual street, the random walkers persistently move along the street until they reach the next street that intersects with the current one. As soon as the random walkers reach the intersection, a decision as to which next street has to make, and then move towards the next intersection, so on and so forth. The random behaviour is obviously occurred at the topological level rather than at the geometric level.