Street Network Generation with Adjustable Complexity Using k-Means Clustering

The transportation system is an important part of the daily lives of a major portion of the world’s population. Therefore, innovative applications are designed to improve the experience of drivers and pedestrians in this system as technological advances allow. These applications analyze street networks to come up with suggestions such as fastest route or optimal ride-sharing route. Graph theory has been used frequently for these analyses, where each junction is represented by a vertex and each road segment is represented by an edge. However, applications also require grouping of these vertices so that they can run optimization methods in a larger scale.In this paper, we propose a method for creating street networks with adjustable complexity. By using k-Means clustering, our mechanism allows for the increase or decrease of the gradient of a street network. The implementation results demonstrate the proposed method’s efficiency and flexibility for providing street network graphs with adjustable complexity.