Parallel best-first branch-and-bound in discrete optimization A framework.pdfVIP

下载本文档

6
0
约7.95万字
约 30页
2017-04-05 发布于江苏
举报
版权申诉

Parallel best-first branch-and-bound in discrete optimization A framework.pdf

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Parallel best-first branch-and-bound in discrete optimization A framework

Parallel Best-First Branch-and-Bound inDiscrete Optimization: a Framework?Ricardo Corre?a??LMC { IMAG46, av. Felix Viallet, 38031Grenoble Cedex, FranceRicardo.Correa@imag.fr Afonso FerreiraCNRS { LIP { ENS-Lyon46, allee dItalie, 69364Lyon Cedex 07, Franceferreira@lip.ens-lyon.frAbstract. In discrete optimization problems, we search for an optimalsolution among all vectors in a discrete solution space that satisfy a set ofconstraints, and the search eciency is of considerable importance sinceexhaustive search is often impracticable. The method called branch-and-bound (noted BB) is a heuristic tree search algorithm used in this con-text. Its principle lies in successive decompositions of the original prob-lem in smaller disjoint subproblems until an optimal solution is found,and the search avoids visiting some subproblems which are known notto contain an optimal solution. Given that disjoint subproblems can bedecomposed simultaneously and independently, parallel processing hasbeen widely considered as an additional source of improvement in searcheciency, using the set of processors to concurrently decompose severalsubproblems at each iteration.Parallel BB is traditionally considered as an irregular parallel algo-rithm due to the fact that the structure of the search tree is not knownbeforehand. It may result in unnecessary work if a subproblem that doesnot contain an optimal solution is chosen and assigned to a processorto be decomposed. Additionally, the use of a distributed memory paral-lel system, which are composed of a set of connected processors, eachone with its own local memory, incurs a number of overheads, includingcommunication overheads and idle time due to workload imbalance andcontention on common data structures. This paper reviews several spe-cial techniques that have been developed to address these problems ina way that can be protable either to the parallel processing or to theoperations research community members.1 IntroductionThe use of parall