Solving the Point Feature Cartographic Label Placement Problem Using Jaccard Index As a Measure of Labels Intersection

This paper presents a new approach to solve the Point Feature Cartographic Label Placement (PFCLP) problem. The PFCLP is a relevant problem for Geographic Information Systems, where the objective is to position labels on a map avoiding overlaps to improve legibility. The first techniques proposed to solve this problem considered only the presence/absence of over-lapping, while more recent ones introduced the idea of estimating the intensity of such overlap. This work extends previous works by considering the overlap intensity through the Jaccard index. Traditional mathematical models for PFCLP were compared to the proposed strategy showing that the Jaccard index provided the best results in terms of legibility. A Clustering Search (CS) algorithm was also proposed to compare the performance of our strategy on a set of large-sized instances. The experimental results show good solutions for instances with up to 13206 points.