1986. Sensor placement strategies for dynamic waste collection
Invited abstract in session TD-59: Innovations for a Greener City and Sustainable Urban Logistics II, stream Transportation.
Tuesday, 14:30-16:00Room: Liberty 1.14
Authors (first author is the speaker)
| 1. | Dilay Aktas Dejaegere
|
| Institute for Mobility - CIB, KU Leuven | |
| 2. | Pieter Vansteenwegen
|
| Institute for Mobility - CIB, KU Leuven |
Abstract
Waste collection is an essential urban service that needs to balance operational costs and service levels. Traditionally, it relies on heavy vehicles operating on fixed routes and frequencies based on historical data. However, waste generation is highly unpredictable due to various factors. This leads to overflowing bins or operational inefficiencies such as underutilized vehicle capacities or more frequent collections than needed. With the smart city technologies, sensor-equipped trash bins can offer real-time waste data to perform dynamic collection planning. To make the most out of this technology, studies on smart waste collection typically assume full sensor coverage, i.e., sensors providing waste volume updates for all bins. However, installing and maintaining sensors require significant investment and the operational benefits of equipping some bins might not justify these costs. In this work, we explore different sensor placement strategies for a dynamic waste collection system, where only a subset of bins has sensors while others rely on predictions from historical data. The dynamic waste collection decides when to visit each bin and in which order, in a rolling horizon principle. We conduct experiments to evaluate the performance of different sensor placement strategies under dynamic waste collection scheme.
Keywords
- Logistics
- Mathematical Programming
- Vehicle Routing
Status: accepted
Back to the list of papers