<![CDATA[Rapid urbanization and post-disaster recovery in developing cities present significant challenges for municipal solid waste management, escalating operational costs and environmental burdens. The primary logistical bottleneck is often the inefficient placement of intermediate waste infrastructure. This study addresses this gap by proposing and validating a novel hybrid spatial decision support framework for the optimal siting of Temporary Waste Transfer Stations (TWTS) in Meulaboh, Indonesia. The framework integrates a Geographic Information System (GIS) with the Analytic Hierarchy Process (AHP) and a Random Forest (RF) machine learning model. AHP structures the problem using expert knowledge and local regulatory standards to generate an initial suitability map. To overcome AHP's linearity and subjectivity, this map generates pseudo-labeled data to train the RF model, which learns complex, non-linear relationships among spatial factors. The RF model demonstrated exceptional performance with an Area Under the Curve (AUC) of 0.96. The framework evaluated 43 villages (Gampong), identifying specific areas in Meureubo and Johan Pahlawan as top candidates due to favorable land use and proximity to road networks. This hybrid approach offers a robust, transparent, and scalable methodology for post-disaster urban infrastructure planning.]]>
Copyrights © 2026
