{"title":"边划分的模拟退火","authors":"H. Mykhailenko, G. Neglia, F. Huet","doi":"10.1109/INFCOMW.2017.8116352","DOIUrl":null,"url":null,"abstract":"In distributed graph computation, graph partitioning is an important preliminary step, because the computation time can significantly depend on how the graph has been split among the different executors. In this paper, we propose a framework for distributed edge partitioning based on simulated annealing. The framework can be used to optimize a large family of partitioning metrics. We provide sufficient conditions for convergence to the optimum as well as discuss which metrics can be efficiently optimized in a distributed way. We implemented our partitioners in Apache GraphX and performed a preliminary comparison with JA-BE-JA-VC, a state-of-the-art partitioner that inspired our approach. We show that our approach can provide improvements, but further research is required to identify suitable metrics to optimize as well as to design a more efficient exploration phase for our algorithm without sacrificing convergence properties.","PeriodicalId":306731,"journal":{"name":"2017 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Simulated annealing for edge partitioning\",\"authors\":\"H. Mykhailenko, G. Neglia, F. Huet\",\"doi\":\"10.1109/INFCOMW.2017.8116352\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In distributed graph computation, graph partitioning is an important preliminary step, because the computation time can significantly depend on how the graph has been split among the different executors. In this paper, we propose a framework for distributed edge partitioning based on simulated annealing. The framework can be used to optimize a large family of partitioning metrics. We provide sufficient conditions for convergence to the optimum as well as discuss which metrics can be efficiently optimized in a distributed way. We implemented our partitioners in Apache GraphX and performed a preliminary comparison with JA-BE-JA-VC, a state-of-the-art partitioner that inspired our approach. We show that our approach can provide improvements, but further research is required to identify suitable metrics to optimize as well as to design a more efficient exploration phase for our algorithm without sacrificing convergence properties.\",\"PeriodicalId\":306731,\"journal\":{\"name\":\"2017 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)\",\"volume\":\"1 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2017-05-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2017 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/INFCOMW.2017.8116352\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2017 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/INFCOMW.2017.8116352","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
在分布式图计算中,图分区是一个重要的初步步骤,因为计算时间很大程度上取决于图在不同执行器之间的分割方式。本文提出了一种基于模拟退火的分布式边缘划分框架。该框架可用于优化大量分区指标。我们给出了收敛到最优的充分条件,并讨论了哪些指标可以以分布式方式有效地优化。我们在Apache GraphX中实现了分区程序,并与java - be - java - vc进行了初步比较,这是一种最先进的分区程序,启发了我们的方法。我们表明,我们的方法可以提供改进,但需要进一步的研究来确定合适的指标来优化,以及为我们的算法设计一个更有效的探索阶段,而不牺牲收敛性。
In distributed graph computation, graph partitioning is an important preliminary step, because the computation time can significantly depend on how the graph has been split among the different executors. In this paper, we propose a framework for distributed edge partitioning based on simulated annealing. The framework can be used to optimize a large family of partitioning metrics. We provide sufficient conditions for convergence to the optimum as well as discuss which metrics can be efficiently optimized in a distributed way. We implemented our partitioners in Apache GraphX and performed a preliminary comparison with JA-BE-JA-VC, a state-of-the-art partitioner that inspired our approach. We show that our approach can provide improvements, but further research is required to identify suitable metrics to optimize as well as to design a more efficient exploration phase for our algorithm without sacrificing convergence properties.