New Distributed Partitioning Approach for Large Spatial Networks
Hafssa Aggour, Aziz Mabrouk · 2023
In urban areas, geographic spread of spatial networks continually produces massive data. In fact, the analysis of these spatial networks is squarely related to the partitioning of networks into sub-networks in a balanced manner using scalable analysis methods with the aim of minimizing processing time, namely, partitioning parallel and distributed with (Big Data, ML). Our objective, consist to partition in a distributed way a large spatial network into several Subnets. Therefore, we are conducting a comparative study of distributed partitioning algorithms such as Clustering and Hierarchical Clustering methods with the method that we have proposed in another paper [26] allowing to decompose the spatial network into sub-networks taking into account the Euclidean proximity between the nodes and the different Voronoï generators. the comparative study of the three partitioning methods that we have implemented, namely, Clustering, Hierarchical Clustering and our proposed method, have allowed us to deduce that our method is the most reliable and competitive compared to others, in terms of balance partitions as well as in terms of optimizing the partitioning time.