 |
 |
All, I appreciate this discussion of agent-based vs. integration results, though ultimately wouldn't one conclude that the original integration measures of space syntax are more accurate in predicting actual movements than the agent-based versions? I remember watching the agent-based simulation of agents' movements through the village of Gassin and was puzzled as to how they "clung" to the "rim" of the pathways' deformed wheel and made much less use of the pathways that were highly integrated in the integration explication. From what I remember, there was much more agreement in the Tate Gallery study where there seemed to be a good amount of commonality among agents' movements, integration, and actual visitor movement (as recorded by observers following patrons). I've always been puzzled as to why the move, among space syntax researchers, to an interest in agent simulation? I'm not aware of any writings that make the link conceptually as to how agent-based work is an integral part of space-syntax theory, since space syntax offers a "global" picture grounded in spatial configuration, whereas agent-based models generate a "piecemeal" picture grounded in the limited "abilities" of the "agents." I expect there are strengths and weaknesses to both approaches, and I'm wondering if anyone has laid them out (as well as a history of how agent-based models arose in space syntax). Interesting messages on this matter. Thanks, Farida, for asking the original question. David Seamon