Design and Development of Settlement Spatial Dynamics Model Using Dynamic and Neural Network Systems

Abstract

Abstract The general objective of this study is to project the distribution of settlement land through the development and application of spatial dynamics models using dynamic system methods and neural networks, while the specific objectives are: 1) Compiling spatial data on land characteristics and evaluating settlement land in Maros Regency; 2) Multitemporal mapping of settlement and non-residential land in 2010, 2016 and 2017; 3) Compile Causal loop diagrams and model simulations to determine the dynamics of population projections and the size of residential land needs until 2038; 4) Synchronizing land requirements with the allocation of space utilization in Maros Regency Spatial Planning. 5) Design and build a spatial model to predict the development and direction of settlement land using change prediction of Artificial Neural Network (ANN); and map the projected spatial distribution model of residential land every 5 (five) years. This research conducted a recent approach to multi-temporal observation and spatial optimization studies by integrating dynamic systems with neural network methods in solving settlement land management problems. The approach with the population growth projection model relates to the spatial dynamics of residential land use using the neural network to facilitate policy makers in decision making for settlement land management. The method used consists of 5 stages: stage I. analysis of land characteristics and land use of multi-temporal settlements, through land surveys, laboratory analysis, and spatial characteristics of land; then the land suitability analysis uses matching method; Phase II: mapping multi-temporal land use in 2010, 2016 and 2017 using analysis and extraction of satellite image information. The image used is Landsat medium resolution satellite imagery; stage III: arranging units and interactions and behaviour between units in a system of population growth dynamics, then making causal loop diagrams and then implementing dynamic software systems to simulate population growth used and residential land needs every five years; stage IV: testing the synchronization of the allocation of space utilization in the Spatial Palnning document of the District with the settlement land resulting from the projection; stage V: build a residential land use projection model which begins with change analysis, transition potential, and change prediction with the neural network method; then the projection result of settlement land needs is mapped every five years interval from 2018 - 2038 based on the results of the analysis. The implication of the results of the study will provide an overview of the harmony between the pattern of space and the trend of the development direction of the settlement, so that the resulting method can be used as a basis for revising the Spatial Planning in the future.