بررسی نقش شناخت فضایی در پیش بینی جریان ترافیک شهری از طریق مدل سازی مبتنی بر عامل

Urban systems are highly complex and non-linear in nature, defined by the behaviours and interactions of many individuals. Building on a wealth of new data and advanced simulation methods, conventional research into urban systems seeks to embrace this complexity, measuring and modelling cities with increasingly greater detail and reliability. The practice of transportation modelling, despite recent developments, lags behind these advances. This paper addresses the implications resulting from variations in model design, with a focus on the behaviour and cognition of drivers, demonstrating how different models of choice and experience significantly influence the distribution of traffic. It is demonstrated how conventional models of urban traffic have not fully incorporated many of the important findings from the cognitive science domain, instead often describing actions in terms of individual optimisation. We introduce exploratory agent-based modelling that incorporates representations of behaviour from a more cognitively rich perspective. Specifically, through these simulations, we identify how spatial cognition in respect to route selection and the inclusion of heterogeneity in spatial knowledge significantly impact the spatial extent and volume of traffic flow within a real-world setting. These initial results indicate that individual-level models of spatial cognition can potentially play an important role in predicting urban traffic flow, and that greater heed should be paid to these approaches going forward. The findings from this work hold important lessons in the development of models of transport systems and hold potential implications for policy.