Targeted Intersection Safety in Data-Sparse Cities: A Discrete-Time Microsimulation and Decision-Making Framework Applied to a Hazardous Urban Junction in Libya

Abstract

Urban safety policy choices in low-data contexts frequently face both high uncertainty and pressing timelines. This paper suggests a lightweight and reproducible framework that combines a discrete-time microsimulation of a single, high-risk four-leg intersection with multi-criteria decision-making. In the model, vehicles and pedestrians are generated via Poisson arrival processes, while behavioral variability such as speeding, red-light running, jaywalking, and driver yielding is modeled using probabilistic parameters calibrated to local Libyan traffic conditions. Four low-cost interventions (i.e., speed bumps, a red-light camera, an improved pedestrian crosswalk, and a one-lane roundabout) are evaluated against the baseline. Each strategy is evaluated using four main simulation-based criteria (i.e., reduction in accidents, effect on vehicle delay, cost of implementation, and a pedestrian safety measure) that combine near-miss and waiting-time changes. These outputs are fed into a Technique for Order Preference by Similarity to the Ideal Solution (TOPSIS) analysis under safety-first, cost-sensitive, and pedestrian-focused stakeholder perspectives. Outcomes exhibit clear, mechanism-consistent trends. By reporting uncertainty explicitly and providing scenario-dependent rankings, the framework translates limited local data into clear, defensible decision guidance. The contribution is both practical and methodological; i.e., a sparse-data pipeline that urban areas can readily implement to prioritize first-step safety spending at high-risk intersections.