پديد آورنده :
فلسفين، حسين
عنوان :
ارتقاي كارايي تصحيح خطاي كدهاي خطي و بهبود عملكرد الگوريتم كد گشايي LP وفقي از طريق مقابله با اثر نا مطلوب ساختارهاي مسبب خطا
گرايش تحصيلي :
كامپيوتر- نرم افزار
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر
صفحه شمار :
چهارده،[165]ص.: مصور، جدول، نمودار
يادداشت :
ص.ع. به فارسي و انگليسي
استاد راهنما :
رسول موسوي
استاد مشاور :
محمدسعيد صباغ
توصيفگر ها :
شبه كد كلمات ناصحيح , مجموعه هاي توقف , مجموعه هاي تله پله , پديده كف نرخ خطا , الگوريت هاي كد گشايي مبتني بر برنامه ريزي خطي , الگوريتم هاي كد گشايي تكراري عبور پيام
استاد داور :
محمدعلي خسروي فرد، محمد دخيل عليان، محمدرضا رفسنجاني صادقي
تاريخ ورود اطلاعات :
1395/08/29
دانشكده :
مهندسي برق و كامپيوتر
كد ايرانداك :
ID943 دكتري
چكيده انگليسي :
Improving the error correcting capability of linear codes and enhancingthe performance of adaptive LP decoding algorithm by counteracting the impairing effect of error prone structures Hossein Falsafain h falsafain@ec iut ac ir October 25 2016 Department of Electrical and Computer Engineering Isfahan University of Technology 84156 83111 Isfahan IranSupervisor Dr Sayyed Rasoul Mousavi srm@aut ac irAdvisor Dr Mohammad S Sabbagh sabbagh@cc iut ac irDepartment Graduate Program Coordinator Dr Mohammad Reza Taban Department of Computer Engineering and Information Technology Amirkabir University of Technol ogy 15916 34311 Tehran Iran Department of Industrial Engineering Isfahan University of Technology 84156 83111 Isfahan IranAbstractThe suboptimal behaviour of iterative message passing and linear programming based algorithms fordecoding binary linear codes is attributed to the presence of certain non codeword states into which thedecoding procedures may be trapped The problem of convergence toward these error prone structurescan be alleviated by either modifying the decoding algorithm or improving the structural properties ofthe parity check matrix specifying the code The rst main aim of the present thesis is to describe a novelinteger linear programming based approach for nding a redundant non redundant parity check equa tion whose addition to a parity check matrix can lead to an effective elimination of the stopping sets Oursecond major purpose is to improve the error correcting capability of the adaptive linear programmingdecoding algorithm using a subroutine for on the y i e in an on demand basis during run time gener ation of cut inducing redundant parity check equations Finally we describe a novel branch and boundprocedure for exhaustively enumerating the elementary trapping sets in an arbitrary given Tanner graph The knowledge of the of elementary trapping sets of the given graph can be employed for predicting itserror oor behaviour or for custom tailoring the decoder so that it can get out of a non codeword statein which it could get trapped KeywordsLinear programming decoding elementary trapping sets error rate oor phenomenon iterative message passing decoding non integral pseudo codewords stopping sets IntroductionIt has long been known that the maximum likelihood ML decoding problem of binary linearcodes is N P hard in general This means that there are no ef cient i e with a polynomial time complexity with respect to the code length algorithms known to date to solve the generalML decoding problem and that it is highly likely that no such algorithms exist Does thismean that we should give up hope stop trying and declare that the ef cient transmission ofinformation at low error probability is an impracticable task Not at all Actually the aforemen tioned fact only indicates that no ef cient algorithm is known which solves the ML decodingproblem for all codes But it leaves open the possibility that some subclasses of codes have an
استاد راهنما :
رسول موسوي
استاد مشاور :
محمدسعيد صباغ
استاد داور :
محمدعلي خسروي فرد، محمد دخيل عليان، محمدرضا رفسنجاني صادقي
