Waste collection & street-sweeping route optimization using a 2-stage cluster algorithm & heuristic approaches

Abstract

Waste collection and street-sweeping play a vital role in public health, safety, and overall cleanliness. Since these processes cannot be ignored, they should be done in an efficient manner. The following thesis proposes a novel 2-stage clustering approach, namely the Static and Dynamic Clustering, to divide a municipalities road network into several operational areas in which the routes can be assigned. A method of generating optimal routes within the respective operational areas is also developed so statistics can be used to quantify the improvements made using the proposed clustering methods. The proposed algorithms were used to optimize the waste collection and street-sweeping processes in The City of Oshawa. The results of this work show that the proposed clustering algorithms can generate operational areas that better distribute the workload and overall simulated statistics when compared to existing configurations. Additionally, the proposed techniques may be applied to other routing applications, and other areas of research involving optimizing data partitions using clustering methods, such as machine learning.