ANALYSIS OF REAL RELATIVE ASYMMETRY IN URBAN TRANSPORTATION NETWORK PROBLEMS USING SPACE SYNTAX, REDS, AND MACHINE LEARNING CONCEPTS

Abstract

In the context of urban growth and increasing population density, urban transportation networks face significant challenges such as traffic congestion, infrastructure limitations, and traffic law violations. This study integrates three analytical approaches—Space Syntax, Resolving Efficient Dominating Set (REDS), and Graph Neural Networks (GNN)—to identify strategic locations for the deployment of mobile Electronic Traffic Law Enforcement (ETLE) units and to forecast potential traffic violations. The research focuses on Malang City, Indonesia, and utilizes spatial data and ETLE violation records. Results show that Laksamana Martadinata Street, which has the lowest Real Relative Asymmetry (RRA) value, is a key strategic location for monitoring. The REDS analysis yields a resolving efficient domination number that determines the optimal quantity and placement of mobile ETLE units. GNN-based multi-step time series forecasting successfully predicts traffic violation trends across 29 road segments with a Mean Squared Error (MSE) equal to 0.0173. The novelty of this research lies in the integration of spatial configuration analysis, graph theoretical optimization, and machine learning-based forecasting, offering a comprehensive approach not previously combined in related studies. However, limitations include the use of a single urban case study and constraints in the availability and granularity of violation data, which may affect the generalizability of the findings.