استنتاج ساختار شهری به شکل کارآمد توسط پویایی جریان ترافیک با استفاده از آتاماتای یادگیری

An optimal connection between important areas of cities is highly important in urban planning. This kind of planning is assumed as an optimization problem, where latent links will be discovered solving it. It can take advantage of the traffic flow. The time varying traffic dynamics and the current structure of a city can contain valuable information. Formulating the problem to use the information is very critical. After formulation, two LA-based algorithms are presented in this paper to infer optimal structure: an approach based on distributed learning automata (DLA), and another based on cellular learning automata. The DLA-based algorithm solves the problem by a collectively cooperating network of automata. The CLA-based algorithm addresses the problem as a computational task and solves it using a lattice of cells working together. Favorite structure of a city or optimal structure is estimated utilizing two signals from the environment. We need a test bed to evaluate the performance of the algorithms, therefore synthetic and real data are utilized. The CLA-based method outperforms others in most cases comparing the fitness value. The result is an optimal connectivity matrix which contains the current urban structure and some new necessary links.