Two-stage routing with optimized guided search and greedy algorithm on proximity graph
作者：
Highlights：
• A tailored two-stage routing framework is designed for proximity graph algorithms.
• Optimized guided search significantly accelerates routing in the first stage to better meet the efficiency requirement.
• Optimized greedy algorithm comprehensively detects the results in the second stage to better meet the accuracy requirement.
• Theoretical and experimental analysis demonstrates the state-of-the-art power of the proposed strategy.
• The scalability of the proposed strategy is robust and can be embedded to various proximity graphs.
摘要
•A tailored two-stage routing framework is designed for proximity graph algorithms.•Optimized guided search significantly accelerates routing in the first stage to better meet the efficiency requirement.•Optimized greedy algorithm comprehensively detects the results in the second stage to better meet the accuracy requirement.•Theoretical and experimental analysis demonstrates the state-of-the-art power of the proposed strategy.•The scalability of the proposed strategy is robust and can be embedded to various proximity graphs.
论文关键词：Approximate nearest neighbor search,Proximity graph,Optimized greedy algorithm,Optimized guided search,Two-stage routing strategy
论文评审过程：Received 27 July 2020, Revised 8 May 2021, Accepted 12 July 2021, Available online 14 July 2021, Version of Record 27 July 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.107305
