Optimizing Municipal Solid Waste Collection with Sensor Data

Abstract

This study is about optimizing the municipal solid waste collection: collection scheduling and garbage truck routing. We argue that the use of sensors for monitoring the amount of waste in containers can improve the prediction of accumulation levels and make scheduling more efficient: fewer overflows and fewer unnecessary visits. The benefits that such optimization can bring about are substantial: a reduction in air pollution and traffic and a decrease in operational costs. At the same time, these changes imply that each container will not have a fixed collection frequency anymore but will be collected as late as possible without letting it overflow. Dynamic scheduling will inevitably require dynamic routing: the routes will be defined based on the set of containers chosen for the given date. We will discuss the benefits and the potential drawbacks that these floating schedule and routing may bring about. We approach the problem from a computational and algorithmic perspective and use methods from the fields of combinatorial optimization and operations research to solve the problem. We review some of the exact and heuristic methods and draw our conclusion based on the literature. Finally, we develop and present a minimalistic software kit that consists of an application for receiving and storing sensor data and a QGIS plugin for scheduling and routing.