Tree decomposition-based indexing for efficient shortest path and nearest neighbors query answering on graphs

Abstract

Please cite this article in press as: F. Wei-Kleiner, Tree decomposition-based indexing for efficient shortest path and nearest neighbors query answering on graphs, J. Comput. Syst. Sci. (2015), http://dx. Highlights• We propose an indexing scheme for efficient graph query processing.• We introduce a decomposition algorithm which works for all graphs.• We design query answering algorithms for shortest path and kNN. AbstractWe propose TEDI, an indexing for solving shortest path, and k Nearest Neighbors (kNN) problems. TEDI is based on the tree decomposition methodology. The graph is first decomposed into a tree in which the node contains vertices. The shortest paths are stored in such nodes. These local shortest paths together with the tree structure constitute the index of the graph.Based on this index, algorithms can be executed to solve the shortest path, as well as the kNN problem more efficiently. Our experimental results show that TEDI offers orders-of-magnitude performance improvement over existing approaches on the index construction time, the index size and the query answering.

Optimasi Rute Pengangkutan Sampah dengan Menggunakan Metode Nearest Neighbour (Studi Kasus: Kabupaten Manokwari, Papua Barat)

“…Berbagai macam metode optimasi sistem pengangkutan telah dilakukan oleh peneliti sebelumnya, antara lain menggunakan metode yang berdasarkan sistem informasi geografis (Hidayat, 2013;Ridha et al, 2016), metode tabu search. Metode nearest neighbour sendiri telah banyak diaplikasikan untuk menyelesaikan permasalahan pengangkutan sampah, baik untuk kasus di luar negeri (Du & He, 2012;Li et al, 2019;Wei-Kleiner, 2016;Yu et al, 2016), maupun di Indonesia (Hermawan, 2018;Rahma et al, 2020).…”

Incremental 2-D Nearest-Point Search with Evenly Populated Strips

“…Let us assume a set S of points in a space M and a query point p  M. The nearest-point search (NPS) aims to find the point in S being the closest to p. In its most common form, M is a d-dimensional vector space (2-D in this paper), points correspond to their position vectors, and closeness is expressed with Euclidean distance. Applications can be found in a variety of domains, such as computational geometry [2,20], geographic information systems -GIS [18], bioinformatics [21], image and video compression [3, pattern recognition [16], computer vision [12], robot motion planning [14] ], telecommunications [9], and computer graphics [6]. There are a number of versatile generalizations where the distance metric between spatial points is extended to any quantitative measure of similarity between two generic objects, as one may also measure closeness of pairs of polygons, text strings, images, audio sequences etc.…”