Integer Least Squares Estimation: Theory and Algorithms

2015-11-25  Xiaodong Pan Hits:[]

Speaker:Dr. Jinming Wen (CNRS, Laboratoire LIP (U. Lyon, CNRS, ENSL, INRIA, UCBL), France)

Time:Nov. 12, 14:30-15:30, 2015

Room:X2511

Abstract:

Integer leastsquares (ILS) problems, also referred to as closest vector problems, havearisen from many applications such as GPS, wireless communications,cryptanalysis, bioinformatics etc. A general ILS problem is NP-hard. In this talk, we review some theory and algorithms for ILS. We first introduce two theoretical results we recently obtained, which rigorously justify the use of the well-known Lenstra, Lenstra and Lovasz (LLL) reduction as preprocessing for solving ordinary ILS problem. We then introduce some lower bound techniques to reduce the cost of solving the ILS. Finally, we introduce some other types of ILS and some future research problems.

Pre:Geometry , Algebra and Topology Seminar Next:Degree of $L^2$-Alexander torsion for 3-manifolds

Close

Baidu
sogou