Urban growth prediction using neural network coupled agents-based Cellular Automata model for Sriperumbudur Taluk, Tamil Nadu, India

Abstract

Unplanned urbanization would pose serious threats to both environment and mankind. Hence, urban growth model (UGM) becomes mandatory to predict future growth of a city. In the current study, urban growth of Sriperumbudur Taluk, Tamil Nadu, India was predicted using three types of Cellular Automata (CA) model namely Traditional CA (TCA) model, Agents based Cellular Automata (ACA) Model and Neural Network coupled Agents- based Cellular Automata (NNACA) model. The urban maps of the study region for the years 2009, 2013 and 2016 along with the influencing agents of urbanization namely transportation, industries, elevation and also hotspot locations based on the Government policy were used in the modeling. Analytical Hierarchical Process (AHP) technique was adopted to estimate the weights of the agents for suitability map preparation in ACA model. On validating 2016 predicted outputs, NNACA model proved to be the better urban model (kappa coefficient - 0.72) when compared to TCA and ACA models (kappa coefficient - 0.6 each). Shannon's entropy measure revealed that the urbanization is concentrated in the north-east direction and it is predicted to have an urban sprawl area of 157 km2 in 2020 using NNACA model while the observed urbanization is 113 km2 of the area in 2016.