Cluster-Based Optimization for Sustainable Municipal Solid Waste Collection Sectorization

Efficient and sustainable municipal waste collection is a critical challenge in urban management, requiring data-driven strategies that align operational performance with environmental objectives. Despite the increasing use of clustering and routing techniques in waste management, current approaches often fail to fully integrate spatial clustering with optimization models that account for real-world constraints such as vehicle capacities and workload balancing. This paper addresses that gap by proposing a novel framework that combines clustering algorithms with mixed-integer linear programming (MILP) to optimize the division of a city into sustainable municipal solid waste collection sectors. Using real-world data from Tarnów, Poland, we apply the k-means algorithm to generate spatially and quantitatively balanced clusters of over 10,000 collection points. These clusters serve as input to a MILP model that solves the Sustainable Sectorization of Municipal Solid Waste Collection Problem (SSMSWCP), aiming to minimize disparities in either route lengths or collected waste volumes across a heterogeneous vehicle fleet, including diesel and electric trucks. The proposed method supports strategic decision-making by enabling planners to evaluate alternative sector layouts, fleet configurations, and environmental trade-offs. Computational experiments demonstrate the feasibility and flexibility of the approach and highlight how operational constraints and balancing criteria affect service equity and environmental outcomes. The entire methodology is implemented in open-source software and is transferable to other municipalities facing similar planning challenges.