Within cluster pattern identification: A new approach for optimizing recycle point distribution to support policy implementation on waste management in Malaysia.

Abstract

Despite the government's policies and objectives, Malaysia lags behind in sustainable waste management techniques, particularly recycling. Bins should be located conveniently to encourage recycling and reduce waste. The current model of bin location-allocation is mostly determined by distance. However, it has been identified that previous studies excluded an important factor: litter pattern identification. Litter pattern is important to identify waste generation hotspots and litter distribution. Thus, we proposed the within cluster pattern identification (WCPI) approach to optimize the recycle point distribution. WCPI gathers the information on litter distribution using geotagged images and analyses the pattern distribution. The optimal location for recycle bin can be identified by incorporating k-means clustering to the pattern distribution. Since k-means faces the non-deterministic polynomial-time-hard challenge of selecting the ideal cluster and cluster centre, WCPI used the total within-cluster sum of square on top of k-means clustering. The proposed location by WCPI is validated in terms of accessibility and suitability. Furthermore, this study provides further analysis of carbon footprint that can be reduced by simulating the data from geotagged images. The results show that 10,323.55 kg of carbon emission can be reduced if the litter is sent for recycling. Thus, we believe that locating bins at an optimal location will embark on consumer motivation to dispose of recycled waste, reduce litter and lessen the carbon footprint. At the same time, these efforts could transform Malaysia into a clean and sustainable nation that aims to achieve Agenda 2030.