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

Abstract

The incremental nearest-point search successively inserts query points into the space partition data structure, and the nearest point for each of them is simultaneously found among the previously inserted ones. The paper introduces a new approach which solves this task in 2-D space in a nearly optimal manner. The proposed dynamic partition into parallel strips, each containing a limited number of points structured in the deterministic skip list, successfully prevents situations with over-populated strips, while its further advanced version with two perpendicular partitions and four categories of deterministic skip lists efficiently decreases the number of strips to be examined in a great majority of practical cases.Povzetek: V članku je predstavljen algoritem za reševanje inkrementalnega problema najbližje točke, ki z dinamično delitvijo ravnine v vzporedne trakove preprečuje prenaseljenost le-teh, z dodatno delitvijo, pravokotno na prvo, pa se večinoma izogne tudi preiskovanju prevelikega števila trakov.