Optimizing waste pre-collection model in Indonesian communities: comparing drop-off and door-to-door models through social cost–benefit analysis

Abstract

Indonesia lacks an established waste pre-collection system resulting 90% waste ending up in open-dumping site. This study identifies a 35-m radius per drop-off points as the most optimal model. The City of South Tangerang, on 21 residential areas with a population of 63,157, being examine. Utilizing ArcGIS Pro and Python scripting, 12 scenarios were developed, model drop-off radii of 15–45 m and household limits of 4–32 per drop-off point, including combined models. Spatial clustering KD-Tree algorithm and Proximity-based Routing were employed to determine collection placements and walking distances. Heatmap and ANOVA tests evaluated conformity with residential characteristics. Model then organized into eight frequency schemes, utilizing motorcart and handcart systems with 4-day and 2-day pickup schedules for waste bank and in-house compost programs. This optimized model reduces collection time by up to 81% and costs by 50% compared to traditional door-to-door systems, achieving cost reductions of $30–$38 USD per ton and lowering CO₂ emissions. The discussion covers opportunity costs and occupational risk. This approach supports Indonesia’s waste reduction targets and offers policy alignment for urban waste management.