Data Compression. Prof. Ja-Ling Wu. Department of Computer Science and Information Engineering National Taiwan University
|
|
- Stanley Green
- 6 years ago
- Views:
Transcription
1 ata Coresso Prof. Ja-Lg Wu eartet of Couter Scece ad Iforato Egeerg Natoa Tawa Uversty
2 ata Coresso ca be acheved by assgg short descrtos to the ost frequet outcoes of the data source ad og descrto to the ess frequet outcoes. efto: A Source code C for a rado varabe s a ag fro the rage to * the set of fte egth strgs of sybos fro a -ary ahabet. Let c() deote the codeword corresodg to ad et () deote the egth of c(). E: C(Red) = 00 C(Bue) = s a source code for = {Red Bue} wth ahabet * = {0 }.
3 efto: The eected egth L(c) of a source code c() for a r.v. wth robabty ass fucto () s gve by L c where () s the egth of the codeword assocated wth. w..o.g.: * = {0 -}. 3
4 efto: A code s sad to be o-sguar f every eeet of the rage of as to a dfferet strg *.e. j c( ) c( j ) No-sguarty suffces for a uabguous descrto of a sge vaue of. But we usuay wsh to sed a sequece of vaues of. I such cases we ca esure decodabty by addg a seca sybo ( coa ) betwee ay two codewords. But ths s a effcet use of the seca sybo; oe ca do better Sef-uctuatg or stataeous codes 4
5 efto: A eteso C* of a code C s a ag fro fte egth strgs of to fte egth strgs of * defed by c( ) = c( )c( ) c( ) where c( )c( ) c( ) dcates cocateato of the corresodg codewords. E: If c( ) = 00 ad c( ) = the c( ) = 00 5
6 efto: A code s caed uquey decodabe f ts eteso s o-sguar. I other words ay ecoded strg a uquey decodabe code has oy oe ossbe source strg roducg t. efto: A coded s caed a ref code or a stataeous code f o codeword s a ref of ay other codeword. A stataeous code ca be decoded wthout referece to the future codewords sce the ed of a codeword s edatey recogzabe. For a stataeous code the sybo ca be decoded as soo as we coe to the ed of the codeword corresodg to t. sef-uctuatg A codes No-sguar codes Uquey decodabe codes Istataeous codes 6
7 KRAFT INEQUALITY: Goa of source codg: costructg stataeous codes of u eected egth to descrbe a gve source. Theore (Kraft equaty) For ay stataeous code (ref code) over a ahabet of sze the codeword egths ust satsfy the equaty Coversey gve a set of codeword egths that satsfy ths equaty there ests a stataeous code wth these word egths. 7
8 Proof: Cosder a -ary tree whch each ode has chdre. Let the braches of the tree rereset the sybos of the codeword. For eae the braches fro the root rereset the ossbe vaues of the frst sybo of the codeword. The each codeword s rereseted by a eaf o the tree. The ath fro the root traces out the sybos of the codeword. 8
9 The ref codto o the codewords es that o codeword s a acestor of ay other codeword o the tree. Hece each codeword eates ts descedats as ossbe codewords. Let a be the egth of the ogest codeword of the set of codewords. Cosder a odes of the tree at eve a. Soe of the are codewords soe are descedats of codewords ad soe are ether. a A codeword at eve has descedats at eve a. Each of these descedat set ust be dsjot. Aso the tota uber of odes these sets ust be ess tha or equa to a. Hece sug over a codewords we have 9
10 or a a whch s the Kraft equaty. Coversey gve ay set of codewords whch satsfy the Kraft quaty we ca aways costruct a -ary tree. Labe the frst ode (ecograhcay) of deth as codeword ad ove ts descedats fro the tree. The abe the frst reag ode of deth as codeword etc. Processg ths way we costruct a ref code wth the secfed. 0
11 0 0 Tota # of odes at eve a = 4 Root Tota # of descedats of ode ode (eve ) at eve a = 4 = ode Leve 0 Leve Leve Leve 3 La = Leve 4
12 Note that: A fte ref code aso satsfes the Kraft equaty. Theore (Eteded Kraft equaty) For ay coutaby fte set of codewords that for a ref code the codeword egths satsfy the eteded Kraft equaty Coversey gve ay satsfyg the eteded Kraft equaty we ca costruct a ref code wth these codeword egths.
13 Proof: Let the -ary ahabet be {0 -}. Cosder the -th codeword y y y. Let 0.y y be the rea y uber gve by the -ary easo O y Ths codeword corresods to the terva O y y y O y y y y the set of a rea uber whose -ary easo begs wth O y y y. Ths s a subterva of the ut terva [0]. j y j j 3
14 By the ref codto those tervas are dsjot. Hece the su of ther egths has to be ess tha or equa to. Ths rove that Just as the fte case we ca reverse the roof to costruct the code for a gve that satsfes the Kraft equaty. Frst reorderg the deg so that. Ths sy assg the tervas order fro the ow ed of the ut terva. 4
15 Ota codes: Questo: How to fd the ref code wth the u eected egth? Fd the set of egths satsfyg the Kraft equaty ad whose eected egth L= s ess tha the eected egth of ay other ref code. Mze L= Over the tegers satsfyg 5
16 We egect the teger costat o ad assue equaty the costrat. Lagrage uter J J wefd og og substtutg ths the costrat yedg ota code egths * og og 0 ad hece to fd 6
17 Ths o-teger chose of codeword egths yeds eected codeword egth * * L og H but sce the ust be tegers we w ot aways be abe to set the codeword egths as eq. (). Istead we shoud choose a set of codeword egths cose to the ota set. 7
18 Theore: The eected egth L of ay stataeous -ary code for a r.v. s greater tha or equa to the etroy H ().e. L H () wth equaty ff. Proof: The dfferece betwee the eected egth ad the etroy ca be wrtte as: L H og og og 8
19 9. a teger for a s og ff.e. wth equaty ff 0 og og og we obta ad Lettg H L c r c r H L C r j j
20 Bouds o the Ota Codeegth Sce og / ay ot equa a teger we roud t u to gve teger word egth assgets og where s thesaest teger. These egths satsfy the Kraft equaty sce Ths choce og of og Mutyg by H codeword egths og og ad sug over L H satsfes we oba 0
21 Theore: Let * * * be the ota codeegths for a source dstrbuto P ad a -ary ahabet ad et L* be the assocated eected egth of the ota code ( L*= * ). The H * L H
22 Fro the recedg theore there s a overhead whch s at ost bt due to the fact that og/ s ot aways a teger. Oe ca reduce the overhead er sybo by sreadg t out over ay sybos: Let s cosder a syste whch we sed a sequece of sybos fro. The sybos are assued to draw..d. accordg to (). We ca cosder these sybos to be a suersybo fro the ahabet. efe L to be the eected codeword egth er ut sybo.e. f ( ) s the egth of the codword assocated wth ( ) the
23 3 H L H H L H H H H H E H E L are ot..d. f cose to the etroy. er syboarbtrary codeegth a eected we ca acheve bock egths usg arge by are..d. Sce
24 Theore: The u eected codeword egth er sybo satsfes: H * H L Moreover f s a statoary stochastc rocess. L * H() where H() s the etroy rate of the rocess. The eected uber of bts er sybo requred to descrbe the rocess. 4
25 Theore: The eected egth uder () of the code assget og satsfes q H q E H q Proof: The eected egth s E og q og q og q q H og og q The crease descrtve coety due to correct forato (dstrbuto) 5
26 We have roved that ay stataeous code ust satsfy the Kraft equaty. The cass of uquey decodabe codes s arger tha the cass of stataeous codes so oe eects to acheve a ower eected codeegth f L s zed over a uquey decodabe codes. I the foowg we rove that the cass of uquey decodabe codes does ot offer ay further ossbtes for the set of codeword egths tha do stataeous codes. 6
27 Theore: The codeword egths of ay uquey decodabe code ust satsfy the Kraft equaty coversey gve a set of codeword egths that satsfy ths equaty t s ossbe to costruct a uquey decodabe code wth these codeword egths. Proof: Cosder C K the K-th eteso of the code.e. the code fored by the cocateato of K reettos of the gve uquey decodabe code C. By the defto of uque decodabty the K-th eteso of the code s o-sguar. Sce there are oy dfferet -ary strgs of egth uque decodabty es that the uber of code sequeces of egth the K-th eteso of the code ust be o greater tha. 7
28 8 Let the codeword egths of the sybos be deoted by (). For the eteso code the egth of the code-sequece s: Let s cosder the Kth ower of the Kraft equaty: k k a k k a K K k K k k k k (or k)
29 where a s the au codeword egth ad a() s the uber of source sequeces k ag to codewords of egth. For uquey decodabe code there s at ost oe sequece ag to each code -sequece ad there are at ost code -sequeces. a k k a a k a k a 9
30 Hece j j k k a Sce ths equaty s true for a k t s true the t as k. Sce (k a ) /k we have j j whch s the Kraft equaty. Coversey gve ay set of satsfyg the Kraft equaty we ca costruct a stataeous code as rescrbed. Sce every stataeous code s uquey decodabe we have aso costructed a uquey decodabe code. 30
31 Coroary: A uquey decodabe code for a fte source ahabet aso satsfes the Kraft equaty. Ay subset of uquey decodabe code s aso uquey decodabe; hece ay fte subset of the fte set of codewords satsfes the Kraft equaty. Hece N N The set of achevabe codeword egths s the sae for uquey decodabe ad stataeous codes 3
32 Huffa Codes. A. Huffa A ethod for the costructo of u redudacy codes Proc. IRE vo. 40; Eae: Code word egth codeword Probabty
33 We eect the ota bary code for to have the ogest codewords assged to the sybos 4 ad 5. Both these egths ust be equa sce otherwse we ca deete a bt fro the oger codeword ad st have a ref code but wth a shorter eected egth. I geera we ca costruct a code whch the two ogest codewords dffer oy the ast bt. 33
34 Eae: If 3 we ay ot have a suffcet uber of sybos so that we ca cobe the at a te. I such a case we add duy sybos to the ed of the set of sybos. The duy sybos have robabty 0 ad are serted to f the tree. Sce at each stage of the reducto the uber of sybos s reduced by - we wat the tota uber of the sybos to be +k(-) where k s the uber of eves the tree. Codeword Probabty uy
35 Otaty of Huffa codes There are ay ota codes: vertg a the bts or echagg two codewords of the sae egth w gve aother ota code. w..o.g. we assue P P P Lea: For ay dstrbuto there ests a ota stataeous code (wth u eected egth) that satsfes the foowg roertes:. If P j > P k the j k. The two ogest codewords have the sae egth. 3. The two ogest codewords dffer oy the ast bt ad corresod to the two east key sybos. 35
36 Proof: If P j > P k the j k Cosder C wth the codewords j ad k of C terchaged. The L C LC But P P k Hece we uct have roerty. j j k P P k j P j P j k P 0 ad sce C k j k P P P j s ota L.Thus C k k tsef j C LC satsfes 0 36
37 If the two ogest codewords are ot of the sae egth the oe ca deete the ast bt of the oger oe reservg the ref roerty ad achevg ower eected codeword egth. Hece the two ogest codewords ust have the sae egth. By roerty the ogest codewords ust beog to the east robabe source sybos. If there s a aa egth codeword wthout a sbg the we ca deete the ast bt of the codeword ad st satsfy the ref roerty. Ths reduces the average codeword egth ad cotradcts the otaty of the code. Hece every aa egth codeword ay ota code has a sbg. 37
38 Now we ca echage the ogest egth codewords so that the two owest robabty source sybos are assocated wth two sbg o the tree. Ths does ot chage the eected egth P. Thus the codewords for the two owest robabty source sybos have aa egth ad agree a but the ast bt. If P P P the there ests a ota code wth - = ad codewords c( - ) ad C( ) dffer oy the ast bt. 38
39 39 For a code C satsfyg the roertes of the above ea we ow defe a erged code C - for - sybos as foows: take the coo ref of the two argest codewords (corresodg to the two east key sybos) ad aot t to a sybo wth robabty P - ad P. A the codewords rea the sae. 0 w w w w w P P w w w P w w w P w w w P C C
40 40 where w deotes a bary codeword ad deotes ts egth. The eected egth of the code C s C L C L
41 Thus the eected egth of the code C dffers fro the eected egth of C - by a fed aout deedet of C -. Thus zg the eected egth L(C ) s equvaet to zg L(C - ). Thus we have reduced the robe to oe wth - sybos ad robabty asses (P P P - + P ). we aga ook for a code whch satsfes the roertes of the ea for there - sybos ad the reduce the robe to fd the ota code for - sybos wth the arorate robabty asses obtaed by ergg the two owest robabtes o the revous erged st. 4
42 Proceedg ths way we fay reduce the robe to two sybos for whch the souto s obvous.e. aot 0 for oe of the sybos ad for the other. Sce we have ataed otaty at every stage the reducto the code costructed for sybos s ota. 4
43 Theore: Huffa codg s ota.e. f C * s the Huffa code ad C s ay other code the L(C * )L(C ). Huffa codg s a greedy agorth that t coaesces ( 合併 ) the two east key sybo at each stage. oca otaty goba otaty 43
Generalized Coding Theorem with Different Source Coding Schemes
Iteratoa Joura o Recet ad Iovato Treds Comutg ad Commucato ISSN: 3-869 Voume: 5 Issue: 6 53 57 Geerazed Codg Theorem wth fferet Source Codg Schemes. K. Sharma Jayee Uversty of Egeerg ad Techoogy Raghogarh
More informationK-Map 1. In contrast, Karnaugh map (K-map) method provides a straightforward procedure for simplifying Boolean functions.
K-Map Lesso Objectves: Eve though Boolea expressos ca be smplfed by algebrac mapulato, such a approach lacks clear regular rules for each succeedg step ad t s dffcult to determe whether the smplest expresso
More informationOn the Techniques for Constructing Even-order Magic Squares using Basic Latin Squares
Iteratoal Joural of Scetfc ad Research Publcatos, Volume, Issue 9, September 0 ISSN 50-353 O the Techques for Costructg Eve-order Magc Squares usg Basc Lat Squares Tomba I. Departmet of Mathematcs, Mapur
More informationLecture6: Lossless Compression Techniques. Conditional Human Codes
ecture6: ossless Compresso Techques Codtoal uma Codes -Cosder statoary dscrete arov process, = { s, s, s } wth codtoal pmfs P P s s wth,, tates o Ps/so.9.5.5 Ps/s.5.8.5 Ps/s.5.5.6 -The margal probabltes
More informationShort Note: Merging Secondary Variables for Geophysical Data Integration
Short Note: Mergg Secodary Varables for Geophyscal Data Itegrato Steve Lyster ad Clayto V. Deutsch Departmet of Cvl & Evrometal Egeerg Uversty of Alberta Abstract Multple secodary data from geophyscal
More informationGeometric Distribution as a Randomization Device: Implemented to the Kuk s Model
It. J. Cotem. Math. Sceces, Vol. 8, 03, o. 5, 43-48 HIKARI Ltd, www.m-hkar.com Geometrc Dstrbuto as a Radomzato Devce: Imlemeted to the Kuk s Model Sarjder Sgh Deartmet of Mathematcs Texas A&M Uversty-Kgsvlle
More informationModule 6. Channel Coding. Version 2 ECE IIT, Kharagpur
Module 6 Chael Codg Lesso 36 Coded Modulato Schemes After readg ths lesso, you wll lear about Trells Code Modulato; Set parttog TCM; Decodg TCM; The modulated waveform a covetoal ucoded carrer modulato
More informationAssignment#4 Due: 5pm on the date stated in the course outline. Hand in to the assignment box on the 3 rd floor of CAB.
MATH Assgmet#4 Due: 5pm o the date stated the course outle. Had to the assgmet box o the 3 rd floor of CAB.. Let deote the umber of teror regos of a covex polygo wth sdes, dvded by all ts dagoals, f o
More informationSixth Edition. Chapter 7 Point Estimation of Parameters and Sampling Distributions Mean Squared Error of an 7-2 Sampling Distributions and
3//06 Appled Statstcs ad Probablty for Egeers Sth Edto Douglas C. Motgomery George C. Ruger Chapter 7 Pot Estmato of Parameters ad Samplg Dstrbutos Copyrght 04 Joh Wley & Sos, Ic. All rghts reserved. 7
More informationA New Mathematical Model for a Redundancy Allocation Problem with Mixing Components Redundant and Choice of Redundancy Strategies
Appled Mathematcal Sceces, Vol, 2007, o 45, 222-2230 A New Mathematcal Model for a Redudacy Allocato Problem wth Mxg Compoets Redudat ad Choce of Redudacy Strateges R Tavakkol-Moghaddam Departmet of Idustral
More informationCharacterization and Construction of Permutation Graphs
Ope Joural of Dscrete Mathematcs, 03, 3, 33-38 http://dxdoorg/036/odm033007 Publshed Ole Jauary 03 (http://wwwscrporg/oural/odm) Characterzato ad Costructo of Permutato Graphs Seero V Geraco, Teofa A Rapaut,
More informationLECTURE 4 QUANTITATIVE MODELS FOR FACILITY LOCATION: SERVICE FACILITY ON A LINE OR ON A PLANE
LECTUE 4 QUANTITATIVE MODELS FO FACILITY LOCATION: SEVICE FACILITY ON A LINE O ON A PLANE Learg objectve 1. To demostrate the quattatve approach to locate faclt o a le ad o a plae 6.10 Locatg Faclt o a
More informationA QoS-based Scheme for Planning and Dimensioning of Optical Label Switched Networks
Ths fu text aer was eer revewed at the drecto of IEEE Commucatos Socety subject matter exerts for ubcato the ICC 27 roceedgs. A QoS-based Scheme for Pag ad Dmesog of Otca Labe Swtched Networks Yasse Khf,
More informationPROJECT ACTIVITY ANALYSIS WITHOUT THE NETWORK MODEL. E. Munapo
ROJEC CIVIY NLYSIS WIHOU HE NEWORK MODEL E. Muapo Departet of Decsos Sceces Uversty of South frca, South frca uape@usa.ac.za BSRC hs paper presets a ew proceure for aalysg a aagg actvty sequeces projects.
More informationDISTRIBUTION VOLTAGE MONITORING AND CONTROL UTILIZING SMART METERS
4 th Iteratoal Coferece o Electrcty Dstrbuto Glasgow, -5 Jue 07 DISTRIBUTION VOLTAGE MONITORING AND CONTROL UTILIZING SMART METERS Yoshhto. KINOSHITA Kazuor. IWABUCHI Yasuyuk. MIYAZAKI Toshba Japa Toshba
More informationDescriptive Statistics
Math 3 Lecture I Descrptve tatstcs Descrptve statstcs are graphcal or umercal methods utlsed to summarze data such a way that mportat features of the sample ca be depcted. tatstcs: tatstcs s cocered wth
More informationAn Efficient Max-Dependency Algorithm for Gene Selection
A Effcet Ma-Deedec Algorth for Gee electo achua Peg awrece Berkele Natoal aborator fe ceces Dvso Uverst of Calfora Berkele CA UA 9470. Eal: heg@lbl.gov Fuhu og Duke Uverst Ceter for Cogtve Neuroscece Durha
More informationOptimization of linear functions on a cyclic permutation Based on the random search
ONHMOD. AN INRNAIONAL QUARRLY JOURNAL 206. Vo. 5. No. 3, 2-26 Optmzato of ear fuctos o a cycc permutato Based o the radom search Igor Grebek, Aexey Baraov, Oga hora, ea Gorbacheva Kharkv Natoa Uversty
More informationInformation Theory and Coding
Iformato heory ad Codg Itroducto What s t all aout? Refereces: C..hao, A Mathematcal heory of Commucato, he Bell ystem echcal Joural, Vol. 7, pp. 379 43, 63 656, July, Octoer, 948. C..hao Commucato the
More informationMultiset Permutations in Lexicographic Order
Webste: www.jetae.com ISSN 2250-2459, ISO 9001:2008 Certfed Joural, Volume 4, Issue 1, Jauary 2014 Multset Permutatos Lexcographc Order Tg Kuo Departmet of Marketg Maagemet, Takmg Uversty of Scece ad Techology,
More informationH(X,Y) = H(X) + H(Y X)
Today s Topics Iformatio Theory Mohamed Hamada oftware gieerig ab The Uiversity of Aizu mail: hamada@u-aizu.ac.jp UR: http://www.u-aizu.ac.jp/~hamada tropy review tropy ad Data Compressio Uiquely decodable
More informationOptimal Resource Allocation Scheme for Satisfying the Data Rate Requirement in Hybrid Network of D2D-Cellular
8 JOURAL OF COPUTERS, VOL 9, O 5, AY 4 Otal Resource Allocato Schee for Satsfyg the Data Rate Requreet Hybrd etwor of DD-Cellular Wewe Lu, Yag Yag, Tao Peg, Webo Wag Wreless Sgal Processg ad etwor Lab,
More informationSIMPLE RANDOM SAMPLING
UIT IMPL RADOM AMPLIG mple Radom amplg tructure. Itroducto Obectves. Methods of electo of a ample Lottery Method Radom umber Method Computer Radom umber Geerato Method.3 Propertes of mple Radom amplg Merts
More informationDistributed Online Matching Algorithm For Multi-Path Planning of Mobile Robots
Proect Paper for 6.854 embers: Seugkook Yu (yusk@mt.edu) Sooho Park (dreameo@mt.edu) Dstrbuted Ole atchg Algorthm For ult-path Plag of oble Robots 1. Itroducto Curretly, we are workg o moble robots whch
More informationK-sorted Permutations with Weakly Restricted Displacements
K-sorted Permutatos wth Weakly Restrcted Dsplacemets Tg Kuo Departmet of Marketg Maagemet, Takmg Uversty of Scece ad Techology Tape 5, Tawa, ROC tkuo@takmg.edu.tw Receved February 0; Revsed 5 Aprl 0 ;
More informationPerformance Comparison of Two Inner Coding Structures in Concatenated Codes for Frequency-Hopping Spread Spectrum Multiple-Access Communications
Iteratoal Joural o Recet ad Iovato Treds Computg ad Commucato IN: 31-8169 Volume: 3 Issue: 741-745 erformace Comparso of Two Ier Codg tructures Cocateated Codes for Frequecy-Hoppg pread pectrum Multple-Access
More informationA New Aggregation Policy for RSS Services
A New Aggregato Polcy for RSS Servces Youg Geu Ha Sag Ho Lee Jae Hw Km Yaggo Km 2 School of Computg, Soogsl Uversty Seoul, Korea {youggeu,shlee99,oassdle}@gmal.com 2 Dept. of Computer ad Iformato Sceces,
More informationTime-Frequency Entropy Analysis of Arc Signal in Non-Stationary Submerged Arc Welding
Egeerg, 211, 3, 15-19 do:1.4236/eg.211.3213 Publshed Ole February 211 (http://www.scrp.org/joural/eg) Tme-Frequecy Etropy Aalyss of Arc Sgal o-statoary Submerged Arc Weldg Abstract Kuafag He 1, Swe Xao
More information3D Face Recognition using Longitudinal Section and Transection
Proc. VIIth gta Iage Coutg: Techques ad Acatos, Su C., Tabot H., Ourse S. ad Adraase T. Eds., - ec., Sde Face Recogto usg Logtuda Secto ad Trasecto Yeugha Lee - ae-dog Kogsa-s Kogboo SEECS Yeugaa Uv. 7-79,
More informationDYNAMIC BROADCAST SCHEDULING IN ASYMMETRIC COMMUNICATION SYSTEMS: PUSH AND PULL DATA BASED ON SCHEDULING INDEX AND OPTIMAL CUT-OFF POINT YUFEI GUO
DYNAMIC BROADCAST SCHEDULING IN ASYMMETRIC COMMUNICATION SYSTEMS: PUSH AND PULL DATA BASED ON SCHEDULING INDEX AND OPTIMAL CUT-OFF POINT by YUFEI GUO Preseted to the Faculty of the Graduate School of The
More informationNew Difference Estimator in Two-phase Sampling using Arbitrary Probabilities
IN 1684-8403 Joural o tatstcs olume 15, 008,. 7-16 Abstract New Derece Estmator Two-hase amlg usg Arbtrar Probabltes Asa Kamal 1 ad Muhammad Qaser hahbaz A ew derece estmator has bee costructed two-hase
More informationStatic games: Coordination and Nash equilibrium
Statc game: Coordato ad Nah equlbrum Lecture Game Theory Fall 204, Lecture 2 3.0.204 Dael Spro, ECON3200/4200, Lecture 2 Ratoalzablty about Narrowg dow the belef I have ad the other player may have by
More informationOPTIMAL BUS DISPATCHING POLICY UNDER VARIABLE DEMAND OVER TIME AND ROUTE LENGTH
OPTIMAL BUS DISPATCHING POLICY UNDER VARIABLE DEMAND OVER TIME AND ROUTE LENGTH Amal S. Kumarage, Professor of Cvl Egeerg, Uversty of Moratuwa, Sr Laka H.A.C. Perera, Cetral Egeerg Cosultacy Bureau, Sr
More informationFault-Tolerant Small Cells Locations Planning in 4G/5G Heterogeneous Wireless Networks
1 Fault-Tolerat Small Cells Locatos Plag G/G Heterogeeous Wreless Networks Tamer Omar 1, Zakha Abchar, Ahmed E. Kamal 3, J. Morrs Chag ad Mohammad Aluem 1 Departmet of Techology Systems, East Carola Uversty,
More informationThe optimization of emergency resource-mobilization based on harmony search algorithm
Avalable ole www.ocpr.com Joural of Chemcal ad Pharmaceutcal Research, 04, 6(7):483-487 Research Artcle ISS : 0975-7384 CODE(USA) : JCPRC5 The optmzato of emergecy resource-moblzato based o harmoy search
More informationThermometer-to-binary Encoder with Bubble Error Correction (BEC) Circuit for Flash Analog-to-Digital Converter (FADC)
Thermometer-to-bary Ecoder wth Bubble Error Correcto (BEC) Crcut for Flash Aalog-to-Dgtal Coverter (FADC) Bu Va Heu, Seughyu Beak, Seughwa Cho +, Jogkook Seo ±, Takyeog Ted. Jeog,* Dept. of Electroc Egeerg,
More informationApplied Statistics and Probability for Engineers, 6 th edition December 31, 2013 CHAPTER 6. Section 6-1
Appled Statstcs ad Probablty for Egeers, 6 th edto December 31, 013 CHAPTER 6 Secto 6-1 6-1. No, usually ot. For eample, f the sample s {, 3} the mea s.5 whch s ot a observato the sample. 6-3. No, usually
More informationLong Number Bit-Serial Squarers
Log Number Bt-Seral Squarers E. Chaotaks, P. Kalvas ad K. Z. Pekmestz are th the Natoal Techcal Uversty of Athes, 7 73 Zographou, Athes, Greece. E-mal: lchaot, paraskevas, pekmes@mcrolab.tua.gr Abstract
More informationAn ANOVA-Based GPS Multipath Detection Algorithm Using Multi-Channel Software Receivers
A ANOVA-Based GPS Multpath Detecto Algorthm Usg Mult-Chael Software Recevers M.T. Breema, Y.T. Morto, ad Q. Zhou Dept. of Electrcal ad Computer Egeerg Mam Uversty Oxford, OH 4556 Abstract: We preset a
More informationMulti-target and Fuzzy Cloud computing Resource Scheduling
Advaced Scece ad Techology Letters Vol.111 (NGCIT 15), pp.94-98 http://dx.do.org/1.147/astl.15.111.19 Mult-target ad Fuzzy Cloud coputg Resource Schedulg Yu Zhag School of forato scece ad techology, Zheg
More informationJournal of Engineering Science and Technology Review 10 (3) (2017) Research Article
Jestr Joural of Egeerg Scece ad Techology Revew 0 (3) (07) 73-80 Research Artcle JOURNAL OF Egeerg Scece ad Techology Revew www.jestr.org Cloud-Asssted Cooperatve Mechas for Fle Dowload Moble eer-to-eer
More informationTHE POWER OF LIKELIHOOD RATIO TEST FOR A CHANGE POINT IN BINOMIAL DISTRIBUTION
JAGST Vol. 63 24 Chage pot oal dstruto THE OWER OF LIKELIHOOD RATIO TEST FOR A CHANGE OINT IN BINOMIAL DISTRIBUTION Muda S.M. Gchuh A.W. 2 Khoro J.M.2 Correspodg author Departet of Actuaral & Statstcs
More informationA CONTROL CHART FOR HEAVY TAILED DISTRIBUTIONS. K. Thaga. Department of Statistics University of Botswana, Botswana
A CONTROL CHART FOR HEAVY TAILED DISTRIBUTIONS K. Thaga Departmet of Statstcs Uversty of Botswaa, Botswaa thagak@mopp.ub.bw ABSTRACT Stadard cotrol charts wth cotrol lmts determed by the mea ad stadard
More informationFrequency Assignment for IEEE Wireless Networks
Frequecy Assgmet for IEEE 8 Wreless Networks K K Leug Bell Labs, Lucet Techologes Murray Hll, NJ 7974 k@bell-labscom Byoug-Jo J Km AT&T Labs Research Mddletow, NJ 7748 macsbug@researchattcom Abstract The
More informationOptimal Power Allocation in Zero-forcing MIMO-OFDM Downlink with Multiuser Diversity
Optmal Power Allocato Zero-forcg IO-OFD Dowl wth ultuser Dversty Peter W. C. Cha ad oger S.. Cheg Abstract hs paper cosders the optmal power allocato for a multuser IO-OFD dowl system usg zero-forcg multplexg
More informationChapter 3. Geographical Data Broadcast Cost Models
Chapter Geographcal ata Broadcast Cost Models s dscussed Secto. T s further dvded to to compoets amel Probe Wat ad Bcast Wat. We argue that t mght be more approprate to dvde T to four compoets: Ide-Probe
More informationEfficient Utilization of FlexRay Network Using Parameter Optimization Method
Iteratoal Joural of Egeerg ad Techology, Vol. 8, No. 6, December 2016 Effcet Utlzato of FlexRay Network Usg Parameter Optmzato Method Y. X. Wag, Y. H. Xu, ad Y. N. Xu Abstract FlexRay s a hgh rate of bus
More informationVoltage Contingency Ranking for IEEE 39-Bus System using Newton- Raphson Method
WSEAS TRANSACTIONS o OWER SSTEMS Haer m, Asma Meddeb, Souad Chebb oltage Cotgecy Rag for IEEE 39-Bus System usg Newto- Raphso Method HAER MII, ASMA MEDDEB ad SOUAD CHEBBI Natoal Hgh School of Egeers of
More informationGeneration Reliability Evaluation in Deregulated Power Systems Using Game Theory and Neural Networks
Smart Grd ad Reewable Eergy, 212, 3, 89-95 http://dx.do.org/1.4236/sgre.212.3213 Publshed Ole May 212 (http://www.scrp.org/joural/sgre) 1 Geerato Relablty Evaluato Deregulated Power Systems Usg Game Theory
More informationRedundancy-Allocation in Pharmaceutical Plant Deepika Garg*, Kuldeep Kumar**,G.L.Pahuja***
Deepka Garg et. al. / Iteratoal Joural of Egeerg Scece ad Techology Vol. 2(5), 200, 088-097 Redudacy-Allocato Pharmaceutcal Plat Deepka Garg*, Kuldeep Kumar**,G.L.Pahua*** *RESEARCH SCHOLAR,DEPT. OF MATHEMATICS,N.I.T.,KURUKSHETRA
More informationInfinite Series Forms of Double Integrals
Iteratoal Joural of Data Evelopmet Aalyss ad *Operatos Research*, 4, Vol., No., 6- Avalable ole at http://pubs.scepub.com/jdeaor/// Scece ad Educato Publshg DOI:.69/jdeaor--- Ifte Seres Forms of Double
More information4.3 COLLEGE ALGEBRA. Logarithms. Logarithms. Logarithms 11/5/2015. Logarithmic Functions
0 TH EDITION COLLEGE ALGEBRA 4. Logarithic Fuctios Logarithic Equatios Logarithic Fuctios Properties of LIAL HORNSBY SCHNEIDER 4. - 4. - The previous sectio dealt with epoetial fuctios of the for y = a
More informationAdaptive antenna selection for energy-efficient MIMO-OFDM wireless systems
Uversty of Wollogog Research Ole Faculty of Egeerg ad Iforato Sceces - Papers: Part A Faculty of Egeerg ad Iforato Sceces 04 Adaptve atea selecto for eergy-effcet MIMO-OFDM wreless systes Ngoc Phuc Le
More informationFault-Tolerant Small Cells Locations Planning in 4G/5G Heterogeneous Wireless Networks
Ths artcle has bee accepted for publcato a future ssue of ths joural, but has ot bee fully edted. Cotet may chage pror to fal publcato. Ctato formato: DOI 10.1109/TVT.016.613, IEEE Trasactos o Vehcular
More information5. Random Processes. 5-3 Deterministic and Nondeterministic Random Processes
5. Radom Processes 5- Itroducto 5- Cotuous ad Dscrete Radom Processes 5-3 Determstc ad Nodetermstc Radom Processes 5-4 Statoary ad Nostatoary Radom Processes 5-5 rgodc ad Noergodc Radom Processes 5-6 Measuremet
More informationx y z HD(x, y) + HD(y, z) HD(x, z)
Massachusetts Istitute of Techology Departmet of Electrical Egieerig ad Computer Sciece 6.02 Solutios to Chapter 5 Updated: February 16, 2012 Please sed iformatio about errors or omissios to hari; questios
More informationMOSFET Differential Pair
EEEB73 Electrcs Aalyss & es (6 MOSFET fferetal Par Able t: Lear Oute escrbe the echas by whch a fferetal-e sal a -e sal are pruce a MOSFET fferetal-aplfer. escrbe the c trasfer characterstcs f a MOSFET
More informationA Participation Incentive Market Mechanism for Allocating Heterogeneous Network Services
A Partcpato Icetve Maret Mechasm for Allocatg Heterogeeous Networ Servces Juog-S Lee ad Boleslaw K. Szymas * Noa Research Ceter, Palo Alto, CA 94304 USA * Resselaer Polytechc Isttute, Troy, NY 280, USA
More informationOPTIMAL DG PLACEMENT FOR MAXIMUM LOSS REDUCTION IN RADIAL DISTRIBUTION SYSTEM USING ABC ALGORITHM
teratoal Joural of Revews Computg 009-00 JRC & LLS. All rghts reserved. JRC SSN: 076-338 www.jrc.org E-SSN: 076-3336 OPTMAL PLACEMENT FOR MAXMUM LOSS REDUCTON N RADAL DSTRBUTON SYSTEM USNG ABC ALGORTHM
More informationReview: Our Approach 2. CSC310 Information Theory
CSC30 Informaton Theory Sam Rowes Lecture 3: Provng the Kraft-McMllan Inequaltes September 8, 6 Revew: Our Approach The study of both compresson and transmsson requres that we abstract data and messages
More informationUtilization Optimization for OVSF Multi-code Assignment in WCDMA Networks
J. C. Che et al. / Asa Joural of Health ad Iforato Sees, Vol., Nos. -4,. 49-65, 7 Utlzato Otzato for OVSF Mult-ode Assget WCDMA Networs JUI-CHI CHEN *, HUI-FUANG NG, AND HSIN-LI LIN Deartet of Couter See
More informationSimulating the Components of Resource Intensity Weight: A Good Proxy for Allocating Health Human Resources Across Inpatients?
Suatg the Cooets of Resouce Itesty Weght: A Good Poxy fo Aocatg Heath Hua Resouces Acoss Iatets? a a a, b Ksaaya Basu, Stéhae L. Paé ad Road Wa a Mcosuato Modeg ad Data Aayss Dvso Aed Reseach ad Aayss
More informationSolution 2 Discussion:
Mobile Couicatios Solutio 2 Discussio: 4..2007 Solutio 2.: Multiplexig For ultiplexig, SDMA, TDMA, FDMA, ad CDMA were preseted i the lecture. a.) What liits the uber of siultaeous users i a TDM/FDM syste
More informationRobust Location Tag Generation from Noisy Location Data for Security Applications
Robust Locato Tag Geerato from Nosy Locato Data for Securty Applcatos D Qu, Da Boeh, Sherma Lo, Per Ege, Staford Uversty BIOGRAPHY D Qu s a Ph.D. caddate Aeroautcs ad Astroautcs workg the Global Postog
More informationChaotic Maps: A Tool to Enhance the Performance of OFDM Systems
Chaotc Maps: A Tool to Ehace the Performace of OFDM Systems E. M. El-Bakary, O. Zahra, S. A. El-Doll, ad F. E. Abd El-Same Departmet of Electrocs ad Electrcal Commucatos Egeerg, Faculty of Electroc Egeerg,
More informationA Two Objective Model for Location-Allocation in a Supply Chain
AmrHosse Nobl, Abolfazl Kazem,Alreza Alejad/ TJMCS Vol. 4 No. 3 (22) 392 4 The Joural of Mathematcs ad Computer Scece Avalable ole at http://www.tjmcs.com The Joural of Mathematcs ad Computer Scece Vol.
More informationDeterministic Random Number Generator Algorithm for Cryptosystem Keys
Iteratoal Joural of Computer ad Iformato Egeerg Determstc Radom Number Geerator Algorthm for Cryptosystem Keys Ad A. Maata, Hamza A. A. Al_Sewad Abstract Oe of the crucal parameters of dgtal cryptographc
More informationCS519K: M ULTIMEDIA SYSTEMS STUDENT PROJECTS DATE ANNOUNCED: OCTOBER 25, 2002 DUE DATE:
CS519K: M ULTIMEDIA SYSTEMS STUDENT PROJECTS DATE ANNOUNCED: OCTOBER 25, 2002 DUE DATE: DECEMBER 5, 2002, 11:59PM Basc gropg: Groups of 2 studets each; (or dvdual 1-member groups) There are 6 projects.
More informationSAIDI MINIMIZATION OF A REMOTE DISTRIBUTION FEEDER. Kai Zou, W. W. L. Keerthipala and S. Perera
SAIDI INIIZATIN F A RETE DISTRIBUTIN FEEDER Ka Zou, W. W.. Keerthpala ad S. Perera Uversty of Wollogog School of Electrcal ad Computer Telecommucato Egeerg Wollogog, NSW 2522, Australa Abstract Dstrbuto
More informationOn the Binomial Coefficients and their Interpolation *
O the Bioial Coefficiets ad their Iterolatio * Leohard Euler Let us rereset the exasio of the ower + x i the followig aer by eas of aroriate characters: + x + x + x + x 3 + etc 3 such that the characters
More informationOptimal Packetization Interval for VoIP Applications Over IEEE Networks
145 Afrca Joural of Iformato ad Commucato Techology, Vol. 2, No. 4, December 2006 Optmal Packetzato Iterval for VoIP Applcatos Over IEEE 802.16 Networks Sheha Perera Departmet of Electrcal & Computer Egeerg
More informationExam. Real-time systems, basic course, CDT315. Grading: Swedish grades: ECTS grades:
6 Exam eal-tme systems, basc course, 35 eacher: amr Isovc Phoe: 0 3 73 Exam durato: 08:30 3:30 Help alloed: Pots: calculator ad a laguage dctoary 48 p Gradg: Sedsh grades: ES grades: Importat formato:
More informationOn the Binomial Coefficients and their Interpolation *
O the Bioial Coefficiets ad their Iterolatio * Leohard Euler Let us rereset the exasio of the ower + x i the followig aer by eas of aroriate characters: + x + x + x + x 3 + etc 3 such that the characters
More informationSimulation of rainfall-runoff process by artificial neural networks and HEC-HMS model (case study Zard river basin)
Proceedgs of The Fourth Iteratoal Ira & Russa Coferece 43 Smulato of rafall-ruoff process by artfcal eural etworks ad HEC-HMS model (case study Zard rver bas Mehrdad Akbarpour MSc. Graguate, Water Structures
More informationEEMD and THT Based Gearbox Fault Detection and Diagnosis
Departmet of Electromechacal Egeerg, Shjazhuag Isttute of Ralway Techology, Shjazhuag, Cha, Emal: Hul68@163.com Abstract A ovel approach to fault detecto ad dagoss of gearbox based o Esemble Emprcal Mode
More informationOn LDPC Code Based Massive Random-Access Scheme for the Gaussian Multiple Access Channel
O LDPC Code Based Massve Radom-Access Scheme for the Gaussa Multple Access Chael Luza Medova, Ato Glebov, Pavel Ryb, ad Alexey Frolov Ist. for Iformato Trasmsso Problems, Moscow, Russa, pryb@tp.ru Moscow
More informationFormulation and Analysis of an Approximate Expression for Voltage Sensitivity in Radial DC Distribution Systems
Eerges 015, 8, 996-9319; do:10.3390/e809996 Artcle OPEN ACCESS eerges ISSN 1996-1073 www.mdp.com/joural/eerges Formulato ad Aalyss of a Approxmate Expresso for Voltage Sestvty Radal DC Dstrbuto Systems
More informationA New Chaos-based Image Cipher Using a Hash Function
A ew Chaos-based Image Cpher Usg a Hash Fucto Chog Fu * School of Computer Scece ad Egeerg ortheaster Uverst Sheag 0004, Cha E-mal: fuchog@mal.eu.e du.c Ou Ba The Geeral Hosptal of Sheag Mltar Commad Sheag
More informationImproved Multi-Channel Blind De-Convolution Algorithm for Linear Convolved Mixture DS/CDMA Signals
Joural of Commucatos Vol. 9, No., October 4 Imroved ult-chael Bld De-Covoluto Algorthm for ear Covolved xture DS/CDA Sgals Hao Cheg, Na Yu, ad Ju u Telecommucato Egeerg aboratory, Chegdu Uversty Natoal
More informationUSING ORTHONORMAL FUNCTIONS IN MODEL PREDICTIVE CONTROL
A. Dubravć, Z. Šehć Koršteje ortoormalh fukcja u predktvom upravljaju a osovu modela USING ORTHONORMAL FUNCTIONS IN MODEL PREDICTIVE CONTROL Amla Dubravć, Zea Šehć ISSN -65 UDC/UDK 68.5.66:57.586 Orgal
More informationProposed an Adaptive Bitrate Algorithm based on Measuring Bandwidth and Video Buffer Occupancy for Providing Smoothly Video Streaming
(IJACSA) Iteratoal Joural of Advaced Computer Scece ad Applcatos, ol. 9, No. 2, 218 Proposed a Adaptve Btrate Algorthm based o Measurg Badwdth ad deo Buffer Occupacy for Provdg Smoothly deo Streamg Saba
More informationAvailable online at (Elixir International Journal) Electrical Engineering. Elixir Elec. Engg. 37 (2011)
3908 Belal Mohammad Kalesar et al./ Elxr Elec. Egg. 37 (0) 3908-395 Avalable ole at www.elxrpublshers.com (Elxr Iteratoal Joural) Electrcal Egeerg Elxr Elec. Egg. 37 (0) 3908-395 Optmal substato placemet
More informationAn Anonymous Multi-Receiver Encryption Based on RSA
Iteratoal Joural of etwork Securty, Vol.15, o.4, PP.307-312, July 2013 307 A Aoymous Mult-Recever Ecrypto Based o RSA Le Har 1, Ch-Che Chag 2,3, ad Hsao-Lg Wu 2 (Correspodg author: Ch-Che Chag) epartmet
More informationAdaptive QoS Control for Real-time Video Communication over Wireless Channels
Adaptve QoS Cotrol for Real-tme Vdeo Commucato over Wreless Chaels Dapeg Wu Y. Thomas Hou Wewu Zhu Zhha He Ya-Q Zhag Abstract Robust trasmsso of real-tme vdeo over wreless chaels s a challegg problem.
More informationCHAPTER-4 WIDE BAND PASS FILTER DESIGN 4.1 INTRODUCTION
CHAPTER-4 WIDE BAND PASS FILTER DESIGN 4. INTRODUCTION The bad pass flters suested last chapter are hav the FBW less tha the 2%. I cotrast of that ths chapter deals wth the des of wde bad pass flter whch
More informationNOISE IN A SPECTRUM ANALYZER. Carlo F.M. Carobbi and Fabio Ferrini Department of Information Engineering University of Florence, Italy
NOISE IN A SPECTRUM ANALYZER by Carlo.M. Carobbi ad abio errii Departet of Iforatio Egieerig Uiversity of lorece, Italy 1. OBJECTIVE The objective is to easure the oise figure of a spectru aalyzer with
More informationProject Scheduling with Two Constrained Resources
Proect Schedulg wth Two Costraed Resources Rog-Hwa Huag, Hsag-L Hsao, Szu-ha Hu, Shh-Tg Syu, ug-a Ho, Rou-We Hsu, Yu-Y Lu Departmet of Busess Admstrato Fu e Catholc Uversty No.50, Chug-Cheg Rd., Hschuag
More informationComparative Testing of Dynamic Parameters DACs with Multi-Tone Signals
17 th Syposu IMEKO TC 4, 3 rd Syposu IMEKO TC 19 ad 15 th IWADC Workshop Istruetato for the ICT Era Sept. 8-10, 010, Kosce, Slovaka Coparatve Testg of Dyac Paraeters DACs wth Mult-Toe Sgals Josef Vedral,
More informationFAULT LOCATION ALGORITHM FOR PRIMARY DISTRIBUTION FEEDERS BASED ON VOLTAGE SAGS
FAULT LOCATION ALGORITHM FOR PRIMARY DITRIBUTION FEEDER BAED ON VOLTAGE AG Rodrgo A. F. Perera, Mlade Kezuovc 2 ad José R.. Matova Uversdade Estadual Paulsta FEI/UNEP Ilha oltera, ão Paulo, Brazl 2 Texas
More informationAutomatic Construction of Semantic Dictionary for Question Categorization
Automatc Costructo of Sematc Dctoary for Questo Categorzato Tayog HAO Departmet of Computer Scece, Cty Uversty of Hog Kog Hog Kog, Cha Xglag NI Departmet of Computer Scece ad Techology, Uversty of Scece
More informationThe Balanced Cross-Layer Design Routing Algorithm in Wireless Sensor Networks Using Fuzzy Logic
Sesors 205, 5, 954-9559; do:0.3390/s508954 Artcle OPEN ACCESS sesors ISSN 424-8220 www.mdp.com/joural/sesors The Balaced Cross-Layer Desg Routg Algorthm Wreless Sesor Networks Usg Fuzzy Logc Ng L *, José-Ferá
More informationChapter 1 Introduction to Stochastic Processes
haper Iroduco o Sochasc rocesses Sochasc rocesses A radom varale s a mappg fuco whch assgs oucomes of a radom eperme o real umers see Fg.. Occurrece of he oucome follows cera proaly dsruo. Therefore a
More informationAn ID-based Proxy Authentication Protocol Supporting Public Key Infrastructure
A ID-based Proxy Authetcato Protocol Supportg Publc Key Ifrastructure Shuh-Pyg Sheh, Shh-I Huag ad Fu-She Ho Departmet of Computer Scece ad Iformato Egeerg, ABSTRACT The advatage of the ID-based authetcato
More informationDistributed Joint AP Grouping and User Association for MU-MIMO Networks
IEEE INFOCOM 2018 - IEEE Coferece o Computer Commucatos Dstrbuted Jot AP Groupg ad User Assocato for MU-MIMO Networks Wagkt Wog S.-H. Gary Cha Departmet of Computer Scece ad Egeerg The Hog Kog Uversty
More informationA Novel Bandwidth Optimization Manager for Vehicle Controller Area Network, CAN, System
A Novel Badwdth Optmzato Maager for Vehcle Cotroller Area Network, CAN, System Y WANG, Z-y YOU, L-hu HUI Guzhou Normal Uversty, Guyag, Guzhou 550001, Cha Abstract Ths paper cosders the badwdth lmtato of
More informationLesson 6: Queues and Markov Chains
lde supportg materal Lesso 6: Queues ad Markov Chas Gova Gambee Queug Theor ad Telecommucatos: Networks ad Applcatos d edto, prger All rghts reserved 3 Queug Theor ad Telecommucatos: Networks ad Applcatos
More informationDistributed Generation Planning Strategy with Load Models in Radial Distribution System
Iteratoal Joural of Computer ad Electrcal Egeerg, Vol., No. 3, August 29 93-863 strbuted Geerato lag Strategy wth oad odels Radal strbuto System eepedra Sgh, eveder Sgh ad K S Verma Abstract Ths paper
More informationA HIGH ACCURACY HIGH THROUGHPUT JITTER TEST SOLUTION ON ATE FOR 3GBPS AND 6GBPS SERIAL-ATA
A HIGH ACCURACY HIGH THROUGHPUT JITTER TEST SOLUTION ON ATE FOR 3GBPS AND 6GBPS SERIAL-ATA Yogqua Fa, Y Ca ad Zeljko Zlc LSI Corporato 0 Amerca Parkway NE, Alletow, Pesylvaa 809 Emal: y.ca@ls.com Departmet
More informationInformation Theory and Communication Optimal Codes
Information Theory and Communication Optimal Codes Ritwik Banerjee rbanerjee@cs.stonybrook.edu c Ritwik Banerjee Information Theory and Communication 1/1 Roadmap Examples and Types of Codes Kraft Inequality
More informationA Matrix Representation of an n-person 0-1 Game and Its 0-1 Tail Algorithm to Find (Strictly) Pure Nash Equilibria
Mathematcs ad Compter Scece 26; (): 5-9 http://www.scecepblshggrop.com/j/mcs do:.648/j.mcs.26.2 A Matrx Represetato of a -Perso - Game ad Its - Tal Algorthm to Fd (Strctly) Pre Nash Eqlbra Day Jag Isttto
More information