Solution approaches for the Smart Waste Collection Routing Problem

Gustavo Souza Marques

Key information

Authors:

Gustavo Souza Marques (Gustavo Souza Marques)

Supervisors:

Tânia Rodrigues Pereira Ramos (Tânia Rodrigues Pereira Ramos)

Published in

January 28, 2021

Abstract

Waste collection companies often face inefficiency in their operations, mostly associated with the highly difficulty in predicting the bins waste accumulation. These fluctuations are linked with the presence of uncertainty in the system, which leads to an excessive travelling kilometer usually to collect a small amount of waste. One way to reduce the uncertainty is to explore electronic devices to access real-time information on the bins fill-level and, thus, optimize the collection routes. This problem is known in the literature as the Smart Waste Collection Routing Problem (SWCRP) and some approaches were already proposed to deal with it (Aguiar et al., in press; Ramos et al., 2018). The SWCRP was approached in both works as a Vehicle Routing Problem with Profits (VRPP) and therefore, the present dissertation aims to improve those works addressing some issues that were not considered beforehand, such as the route balance. Three operational management approaches are introduced to define dynamic routes considering real-time information about the bins fill-levels. In addition to the route balance, another branch is investigated to study how to select the bins to be considered by the model. Besides the option that complies all bins, a pre-selection procedure is explored combining the VRPP model with a heuristic method. The solution method is applied in a real-world case study where a potential increase in the operational profit is achieved by up to 450%. Moreover, it is concluded that the solution method can also allow an improvement of up to 69% in the operations efficiency.