Developing Optimization Tools for Municipal Solid Waste Collection in the Argentine City of Berazategui

Abstract

We apply mathematical and computational techniques to the development of an approach for optimizing the waste collection system of the Argentine city of Berazategui, 26 km south of Buenos Aires. Taking full account of the city’s particular characteristics, our objective is to not only improve the system’s efficiency but also ensure equitable workloads for waste collection truck crews, including both drivers and collectors. The optimization problem is partitioned into three stages. In the first stage, a heuristic constructs structurally simple collection zones that are balanced in terms of collectors’ walking distances. In the second stage, a mixed-integer linear programming model designs a collection truck route for each zone and minimizes its length. In the third and final stage, each truck is assigned to two zones in such a way as to equalize to the extent possible the length of drivers’ working day. Because working-day length is influenced by multiple factors, we formulate this objective as a biobjective optimization problem and solve it by integer linear programming coupled with an iterative algorithm. The city implemented the approach in early 2020, resulting in a markedly more equitable workload distribution and significant fuel savings and maintenance expense for the city. History: This paper was refereed. Funding: This work was partly financed by the Universidad de Buenos Aires Ciencia y Técnica [Grant 20020170100495BA] (Argentina) and Proyectos de Investigación Plurianuales - Consejo Nacional de Investigaciones Científicas y Técnicas [Grant 11220200100084CO (Argentina)]. G. A. Durán is also funded by the Instituto Sistemas Complejos de Ingeniería in Chile [Grants Iniciativa Científica Milenio - Fondo de Innovación para la Competitividad P05-004-F and Comisión Nacional de Investigación Científica y Tecnológica FB0816].