Expected Worst-case Performance of Hash Files
|
|
- Gwendolyn Daniela McKenzie
- 6 years ago
- Views:
Transcription
1 Expected Worst-cse Performnce of Hsh Files Per-Ake Lrson Deprtment of Informtion Processing, Abo Akdemi, Fnriksgtn, SF-00 ABO 0, Finlnd The following problem is studied: consider hshfilend the longest probe sequence tht occurs when retrieving record. How long is this probe sequence expected to be? The pproch tken differs from trditionl worst-cse considertions, which consider only the longest probe sequence of the worst possiblefileinstnce. Three overflow hndling schemes re nlysed: uniform hshing (rndom probing), liner probing nd seprte chining. The numericl results show tht the worst-cse performnce is expected to be quite resonble. Provided tht the hshing functions used re well-behved, extremely long probe sequences re very unlikely to occur. INTRODUCTION It is well-known tht the worst-cse performnce of hsh file is 0(n), where n is the number of records stored. This is true for ny of the bsic opertions: retrievl, insertion or deletion of record. This result is seen immeditely by considering the longest probe sequence of the worst possible file instnce. However, the probbility tht the worst possible file instnce will ctully occur is incredibly smll. It is n extremely pessimistic view which does not give ny informtion bout the expected worst-cse performnce of 'norml, wellbehved' hsh file. In this pper the worst-cse performnce of hsh file is pproched from different ngle. The problem studied cn be stted s follows: consider certin hsh file nd the longest of the probe sequences required to retrieve record stored in the file. Wht is the expected length of this probe sequence? In other words, not only the longest probe sequence of the worst possible file instnce is of interest, but the whole distribution of the longest probe sequence over ll file instnces. The term 'expected worst-cse' will be used to refer to the ltter type of nlysis. The former, trditionl pproch might perhps be clled the worst-of-the-worst pproch. The nlysis reported here is restricted to the retrievl performnce of externlly stored hshfiles. The length of probe sequence, or simply the serch length, is mesured in the number of ccesses to secondry storge. Three different overflow hndling schemes re studied: uniform hshing (rndom probing) (section two), liner probing (section three) nd seprte chining (section four). The bucket size is one or lrger thn one. Only the cse of initil loding is considered, i.e. no deletions re llowed. So fr only two nlyses of the expected worst-cse performnce of hshing hve been reported: one by Gonnet nd one by Lrson. Gonnet studied internlly stored hsh tbles (bucket size one) ssuming tht overflow records re hndled by rndom probing or by seprte chining. He showed tht the expected length of the longest probe sequence grows logrithmiclly in the cse of rndom probing, nd sublogrithmiclly in the cse of seprte chining. Lrson nlysed new hshing scheme clled Liner Hshing with Prtil Expnsions. Gonnet ws ble to obtin simple pproximte formule for the schemes considered. The modelling of the schemes considered in this pper is considerbly more complex mthemticlly, nd therefore we will hve to be content with numericl results. From elementry probbility theory we recll the following useful theorem. In rndom smple x x, x,..., x n the probbility tht the lrgest observed vlue is less thn or equl to fixed vlue x 0 equls the product of the probbilities tht ech observed vlue is less thn or equl to x 0, i.e. P(mx(Xj) < x o ) = P(JCI < *o) P(* ^*o)- Hx n <x 0 ). This is true if ll the rndom vribles re mutully independent. The nlysis of ech scheme follows the sme bsic pttern. First, the probbility distribution of the length of probe sequence (successful serch) is derived. Therefter, the probbility distribution of the length of the longest probe sequence, herefter bbrevited to lips, in smple of n sequences is redily found by pplying the bove theorem. Finlly the expected vlue of the lips is numericlly computed nd tbulted. The results re pproximte becuse the rndom smple is considered to be drwn from n infinite file. This simplifiction ws necessry becuse, for the schemes considered, either the distribution of the length of probe sequence in finite file is not known or, when known, the numericl computtions involved re fr too time-consuming. The results given here overestimte the expected length of the longest probe sequence. For modertely lrge files the error is no more thn few per cent. To fcilitte comprisons the expected lengths of both successful nd unsuccessful serches hve lso been computed nd compiled into n ppendix. Note tht these expected vlues re for n rbitrry successful or unsuccessful serch, not the longest one. UNIFORM HASHING ~ Consider hshfileconsisting of m buckets, ech bucket hving cpcity of b records, b >. There re n records stored in thefile.the storge utiliztion, denoted by, is then = n/(mb). We shll ssume in this section tht overflow records re hndled by uniform hshing (rndom probing, rehshing); i.e. ech overflowing record is rehshed rndomly until non-full bucket is found (see Ref. ). Let P(L = k),k = \,,..., denote the probbility tht retrievl of record requires k ccesses. For finite m this CCC-OOKM6/8/00-07 $0.00 Heyden & Son Ltd, 98 THE COMPUTER JOURNAL, VOL., NO.,98 7 Downloded from
2 P.-A. LARSON distribution is known only for the cse b =. For b >, the symptotic distribution, obtined by letting m oo keeping the storge utiliztion constnt, ws derived by Lrson. It cn be computed s ) P(L = k) = (I/) P b (yf~ '( - Jo where b-l (ybv/jl nd where the function x() is (implicitly) defined by Eq. (). fr- = exp( xb){\ x) ]T (xbf/kl exp( xb)(xby > ~ /(b )! () Let Q k denote the probbility tht retrievl of record requires t most k ccesses, i.e. Q k P(L < k). This probbility is k Q k = I 0/) P^y-'U - P b (y)) dy j=i Jo which cn be reduced to -l-/oof* 0 ] Jo The integrl must be computed numericlly. The vlue x() must lso be computed numericlly from Eqn (). Consider smple of n records from n infinite file with bucket size b nd storge utiliztion. By pplying the theorem mentioned bove, we immeditely find tht the probbility tht the longest probe sequence encountered in the smple is t most k ccesses equls QJJ. The expected length of the longest probe sequence is thus R n = I *(QZ - QZ-,) = i + I (l - QZ) The distribution QZ hs been plotted in Fig. for n = looo.oc = 0.8nd6 =,. We see tht with probbility greter thn 90%, the length of the longest probe sequence will not exceed 9 when b = nd 6 when b =. The length of the longest probe sequence is very stble; the probbility tht it will be either or is s high s 7. per cent when b =. Tble shows the expected length of the longest probe sequence for few prmeter combintions. The vlues re surprisingly low, especilly for lrger buckets. The longest probe sequence in file with b =, = 0.8 nd storing one million records is expected to be only.9 ccesses, for exmple. For expected serch lengths, see the Appendix i i r J b = S 6 8 k (ccesses) Figure. The probbility tht lips <, k for uniform hshing, n = 00, o=0.8. b =,. LINEAR PROBING ~ Liner probing is one of the erliest nd most widely used techniques for hndling overflow records. The serch pth of record hshing to bucket k is k, k +,..., m,0, k nd the record is inserted into the first non-full bucket encountered on this pth. The performnce of liner probing hs been nlysed by severl uthors, but either only the cse b = or only the expected performnce is considered, ' ' 6 ' 7 The first nlysis of" the cse b>, which includes the whole probbility distribution, not merely the expected vlue, of the serch length, ws published in 978 by Blke nd Konheim. 8 The nlysis below is bsed on their model, but the formule re rewritten in form better suited for numericl computtions. Assuming n infinite file, the probbility tht k ccesses will be required to retrieve record cn be computed s {<p b (x)exp(-xb) s=0 The integrl must be computed numericlly. The numbers {V bj+s },j > 0,0 <, s <, b, occurring in the formul do not depend on the storge utiliztion x. They cn be computed using the recurrence V hi+s = j i=l for s = 0 + *)){(/+l)v w+,_, for =,,..., b - Tble. Expected Ups for uniform hshing No. of records Bucket size, b = = = = THE COMPUTER JOURNAL, VOL., NO.,98 Heyden & Son Ltd, 98 Downloded from
3 EXPECTED WORST-CASE PERFORMANCE OF HASH FILES where the boundry vlues for./ = 0 re given by V s = Wef/sl, s = 0,,..., b -. The numbers {V bj+s } re relted to the numbers {T b>bj+s } defined in Ref. 8 by V bj + s = T 6, 6j+s (6/e) b -' + 7(fe/ + s)\. The chnge hs been mde purely for numericl resons; the numbers {T(,,bj+ S } grow extremely rpidly. The function q> b (x) is normliztion fctor which ensures tht the probbilities sum to one. It is given by cp b (x) = b(l - 7= where <o is primitivefethroot of unity; o> = cos(nlb) + isin(7i/fe), for exmple. The complex-vlued function 0(z), zec, is defined by the series r = 0 The series converges slowly, however, nd numericlly the function 6(z) is most esily computed from the functionl eqution 9{z) = exp (zd(z)) using Newton- Rphson itertion. Afirstpproximtion is redily found by computing few terms of the series. The rest is then strightforwrd. The probbility tht retrievl of record requires t most k ccesses is nd the probbility tht the longest probe sequence occurring in smple of size n is of length k or less is Q. The expected lips is then given by R n = + k =i0-qz)- Exctly s for uniform hshing, the nlysis of liner probing is lso pproximte nd slightly overestimtes the expected vlue. The error is thus 'on the sfe side'. Compred with uniform hshing the numericl computtions re considerbly more complex. Figure shows the cumultive probbility distribution QJj for two different bucket sizes. Compred with uniform k (ccesses) Figure. The probbility tht lips < k when using liner probing, n= 00, = 0.8, b =,. hshing, the probbilities grow very slowly; the risk of long probe sequences occurring is much greter. The expected vlues given in Tble re considerbly lrger thn the corresponding vlues for uniform hshing. This poor worst-cse performnce is not unexpected if the tendency of liner probing to crete long clusters of full buckets is borne in mind. For expected serch lengths, see the Appendix. SEPARATE CHAINING When using seprte chining, overflow records re stored by linking one or more secondry pges from seprte storge re to the overflowing (primry) pge. ' 9 Ech secondry pge holds records from only one chin, i.e. chins re not llowed to colesce. The secondry pge size, denoted by c, c >, my differ from the primry pge size b. Consider file where the expected number of records per bucket is zb, z > 0. The prmeter z denotes the lod fctor. Note tht the lod fctor is not the sme s storge utiliztion nd tht the lod fctor my be lrger thn one. For lrge file the probbility tht j,j = 0, records hsh to bucket cn be pproximted by the Poisson probbility exp(-zb)(zb) J /j\. Hence the probbility tht chin of pges will be of length k or less, k >, (the primry pge plus k secondry pges) is = exp(-z6) If file consisting of N primry pges is considered to hve been creted by smpling from n infinite file, the probbility tht the longest chin will be of length k or less is Q k (z) N. This is t the sme time the probbility tht the longest probe sequence will be of length k or fewer ccesses. Observe tht we now hve the number of primry pges in the exponent, not the number of records s in the two previous cses. Seprte chining is totlly different type of overflow hndling scheme from uniform hshing nd liner probing. Performnce figures for seprte chining cnnot be compred with those of the two other techniques unless the following two conditions re fulfilled. () Storge utiliztion is the sme for ll three methods, nd () files of the sme size, i.e. the sme number of records, re considered. These conditions cn be fulfilled by djusting the prmeters z nd N respectively. In order to be ble to compute the storge utiliztion, we must mke two simplifying ssumptions. The spce occupied by pointers, (one per pge) nd overflow spce Tble. Expected Ups for liner probing No. of records Bucket size, b = =»» = " ti = 0.8 " = " Heyden & Son Ltd, 98 THE COMPUTER JOURNAL, VOL., NO.,98 9 Downloded from
4 P.-A. LARSON llocted but not in use, i.e. empty overflow pges, is ignored. They both depend on the detils of the implementtion. Under these ssumptions the lod fctor z required to chieve storge utiliztion cn be computed numericlly from Eqn (). OO zb/ = b + cexp(-zb) (k + ) fc = O t (zbf +kc+i /(b + kc + 0! () This eqution defines z s function of nd this function will be denoted by z(). The expected number of records per bucket is z{d)b. To store n records in thefile, the number of buckets must be pproximtely = n/(z()b). () Finlly the expected length of the longest probe sequence cn be computed s fc=l To sum up: given storge utiliztion nd pge size b nd c, the required lod fctor z() is computed numericlly from Eqn (). Therefter the number of pges N() is given by Eqn () nd finlly the expected vlue cn be computed. Figure shows the cumultive probbility distribution of the lips for = 0.8, n = 00, c = nd b =,,. The computed lod fctor is z(0.8) =. for b=\, z(0.8) = for b = nd z(0.8) = 0.8 for b =. Observe tht seprte chining behves differently from the two other techniques: the probbility of long probe sequences increses with incresing primry pge size. The sme behviour cn be observed in Tble for Figure. chining,/;. Ai o r =H 6 8 k (ccesses) tht llpss k when using seprte r h c in / The probbility n D ft higher storge utiliztions. For expected serch lengths, see the Appendix. The tbles lists numericl results for the cse c = only, the secondry pge size most often used in prctice. CONCLUSIONS The expected worst-cse retrievl performnce of three hshing schemes, uniform hshing (uh), liner probing (lp) nd seprte chining (sc) hs been nlysed. Figure is n ttempt to summrize the mjor results of this nlysis. The nlysis revels tht the expected worst-cse performnce is not very bd. With the exception of liner probing using smll buckets, the longest probe sequence occurring in file is expected to be of quite resonble length. At lest to the uthor, this ws surprise. The expected length of the longest probe sequence increses with the file size. For uniform hshing the growth is logrithmic, nd for the two others nerly logrithmic. It is not esily seen from Fig., but the growth is (/I 8u < 0 Figure. uh, ( ) (b) - - s' /' Number of records /' _s' s' s' s s /' i i i i i i i i i i J Number of record: > J Expected length of the longest probe sequence s Fnr h ' IM Fr-- <> \D) rr > Tble. Expected Ups for seprte chining withc = l No. of records Bucket size, b = = 6 ij = 0.8 " = THE COMPUTER JOURNAL, VOL., NO.,98 Heyden & Son Ltd, 98 Downloded from
5 EXPECTED WORST-CASE PERFORMANCE OF HASH FILES sublogrithmic for seprte chining nd slightly fster thn logrithmic for liner probing. The worst-cse performnce of uniform hshing is considerbly better thn tht of liner probing. This ws to be expected becuse liner probing tends to crete long clusters of full buckets. For smll bucket sizes the performnce of seprte chining is better thn tht of uniform hshing, but for lrge buckets the order is reversed. The min result of ll the mthemtics bove is perhps simply this: it offers theoreticl explntion for the empiricl observtion tht very long probe sequences re unlikely to occur in well-designed hsh file. A few comments on the results listed in the ppendix re in order. The expected serch lengths of uniform hshing is lwys lower thn those of liner probing. This comes s no surprise. However, for lrge buckets the serch lengths for both uniform hshing nd liner probing re lower thn those of seprte chining with c =. The explntion for this somewht surprising fct is quite simple: When the primry pge size is incresed, keeping the totl size of the primry storge re constnt, the totl number of overflow records decreses,, but so does the number of overflow chins. So even though there re, in totl, fewer overflow records, there my be more of themper chin, resulting in longer chins. This problem cn be overcome by using either lrger overflow pges or severl overflow chins per primry pge, selecting one of the chins by hshing. Acknowledgement The help of B. Qvist in solving some of the computtionl problems ws impertive nd is grtefully cknowledged. REFERENCES. G. H. Gonnet, Expected length of the longest probe sequence in hsh code serching. Journl of the ACM 8 (No. ) (98).. P.-A. Lrson, Performnce nlysis of liner hshing with prtil expnsions. ACM Trnsctions on Dtbse Systems (to pper).. D. E. Knuth, The Art of Computer Progrmming, Vol. : Sorting nd Serching. Addison Wesley, Reding, Msschusetts (97).. P.-A. Lrson, Anlysis of uniform hshing. Abo Akdemi, Reports on Computer Science & Mthemtics Ser. A, No. (978).. P.-A. Lrson, Frequency loding nd liner probing, BIT, Bind 9 (No. ), -8(979). 6. G. Schy nd W. G. Spruth, Anlysis of file ddressing method. Communictions of the ACM (No. ), 9-6 (96). 7. G. Tiniter, Addressing for rndom-ccess storge with multiple bucket cpcities. Journl of the ACM (No. ), 07- (96). 8. I. F. Blke nd A. G. Konheim, Big buckets re (re not) better I Journl of the ACM (No. ), (977). 9. J. A. vn der Pool, Optimum storge lloction for initil loding of file. IBM Journl of Reserch nd Development 6 (No. 6), 79-86(97). Received June 98 Heyden & Son Ltd, 98 APPENDIX Expected length of successful nd unsuccessful serches The tbles below list the expected length of n rbitrry (not the longest) successful nd unsuccessful serch for uniform hshing, liner probing nd seprte chining. The results re symptotic, i.e. for n infinite file, nd slightly overestimte the serch lengths for correspondingfinitefile. Tble Al. Expected serch lengths for uniform hshing Bucket size Successful serch Unsuccessful serch ; Ct Heyden & Son Ltd, 98 THE COMPUTER JOURNAL, VOL., NO.,98 Downloded from
6 P.-A. LARSON Tble A. Expected serch lengths for liner probing ' ' 8 Bucket size SSuccessful serch Unsuccessful serch Tble A. Expected serch lengths for seprting chining (c = I) - 6 Bucket size Successful serch Unsuccessful serch ' denotes the storge utiliztion. The corresponding lod fctor is computed from Eqn (). THE COMPUTER JOURNAL, VOL, NO.,98 Heyden & Son Ltd, 98 Downloded from
Parallel Square and Cube Computations
Prllel Squre nd Cube Computtions Albert A. Liddicot nd Michel J. Flynn Computer Systems Lbortory, Deprtment of Electricl Engineering Stnford University Gtes Building 5 Serr Mll, Stnford, CA 945, USA liddicot@stnford.edu
More informationa < a+ x < a+2 x < < a+n x = b, n A i n f(x i ) x. i=1 i=1
Mth 33 Volume Stewrt 5.2 Geometry of integrls. In this section, we will lern how to compute volumes using integrls defined by slice nlysis. First, we recll from Clculus I how to compute res. Given the
More informationFig.1. Let a source of monochromatic light be incident on a slit of finite width a, as shown in Fig. 1.
Answer on Question #5692, Physics, Optics Stte slient fetures of single slit Frunhofer diffrction pttern. The slit is verticl nd illuminted by point source. Also, obtin n expression for intensity distribution
More informationA New Learning Algorithm for the MAXQ Hierarchical Reinforcement Learning Method
A New Lerning Algorithm for the MAXQ Hierrchicl Reinforcement Lerning Method Frzneh Mirzzdeh 1, Bbk Behsz 2, nd Hmid Beigy 1 1 Deprtment of Computer Engineering, Shrif University of Technology, Tehrn,
More information9 Graph Cutting Procedures
9 Grph Cutting Procedures Lst clss we begn looking t how to embed rbitrry metrics into distributions of trees, nd proved the following theorem due to Brtl (1996): Theorem 9.1 (Brtl (1996)) Given metric
More informationPointwise convergence need not behave well with respect to standard properties such as continuity.
Chpter 3 Uniform Convergence Lecture 9 Sequences of functions re of gret importnce in mny res of pure nd pplied mthemtics, nd their properties cn often be studied in the context of metric spces, s in Exmples
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd business. Introducing technology
More informationMidterm 2 Sample solution
Nme: Instructions Midterm 2 Smple solution CMSC 430 Introduction to Compilers Fll 2012 November 28, 2012 This exm contins 9 pges, including this one. Mke sure you hve ll the pges. Write your nme on the
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd business. Introducing technology
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd business. Introducing technology
More information1. SEQUENCES INVOLVING EXPONENTIAL GROWTH (GEOMETRIC SEQUENCES)
Numbers nd Opertions, Algebr, nd Functions 45. SEQUENCES INVOLVING EXPONENTIAL GROWTH (GEOMETRIC SEQUENCES) In sequence of terms involving eponentil growth, which the testing service lso clls geometric
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd business. Introducing technology
More informationChapter 2 Sensitivity Analysis: Differential Calculus of Models
Chpter 2 Sensitivity Anlysis: Differentil Clculus of Models Abstrct Models in remote sensing nd in science nd engineering, in generl re, essentilly, functions of discrete model input prmeters, nd/or functionls
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd processes. Introducing technology
More informationUnit #9 : Definite Integral Properties, Fundamental Theorem of Calculus
Unit #9 : Definite Integrl Properties, Fundmentl Theorem of Clculus Gols: Identify properties of definite integrls Define odd nd even functions, nd reltionship to integrl vlues Introduce the Fundmentl
More information3.5.1 Single slit diffraction
3..1 Single slit diffrction ves pssing through single slit will lso diffrct nd produce n interference pttern. The reson for this is to do with the finite width of the slit. e will consider this lter. Tke
More informationA Transportation Problem Analysed by a New Ranking Method
(IJIRSE) Interntionl Journl of Innovtive Reserch in Science & Engineering ISSN (Online) 7-07 A Trnsporttion Problem Anlysed by New Rnking Method Dr. A. Shy Sudh P. Chinthiy Associte Professor PG Scholr
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd processes. Introducing technology
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd processes. Introducing technology
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd processes. Introducing technology
More informationMATH 25 CLASS 5 NOTES, SEP
MATH 25 CLASS 5 NOTES, SEP 30 2011 Contents 1. A brief diversion: reltively prime numbers 1 2. Lest common multiples 3 3. Finding ll solutions to x + by = c 4 Quick links to definitions/theorems Euclid
More information12-B FRACTIONS AND DECIMALS
-B Frctions nd Decimls. () If ll four integers were negtive, their product would be positive, nd so could not equl one of them. If ll four integers were positive, their product would be much greter thn
More information3.5.1 Single slit diffraction
3.5.1 Single slit diffrction Wves pssing through single slit will lso diffrct nd produce n interference pttern. The reson for this is to do with the finite width of the slit. We will consider this lter.
More information4452 Mathematical Modeling Lecture 4: Lagrange Multipliers
Mth Modeling Lecture 4: Lgrnge Multipliers Pge 4452 Mthemticl Modeling Lecture 4: Lgrnge Multipliers Lgrnge multipliers re high powered mthemticl technique to find the mximum nd minimum of multidimensionl
More informationP(r)dr = probability of generating a random number in the interval dr near r. For this probability idea to make sense we must have
Rndom Numers nd Monte Crlo Methods Rndom Numer Methods The integrtion methods discussed so fr ll re sed upon mking polynomil pproximtions to the integrnd. Another clss of numericl methods relies upon using
More informationUnion-Find Problem. Using Arrays And Chains. A Set As A Tree. Result Of A Find Operation
Union-Find Problem Given set {,,, n} of n elements. Initilly ech element is in different set. ƒ {}, {},, {n} An intermixed sequence of union nd find opertions is performed. A union opertion combines two
More informationApproximation by NURBS with free knots
pproximtion by NURBS with free knots M Rndrinrivony G Brunnett echnicl University of Chemnitz Fculty of Computer Science Computer Grphics nd Visuliztion Strße der Ntionen 6 97 Chemnitz Germny Emil: mhrvo@informtiktu-chemnitzde
More informationEngineer To Engineer Note
Engineer To Engineer Note EE-169 Technicl Notes on using Anlog Devices' DSP components nd development tools Contct our technicl support by phone: (800) ANALOG-D or e-mil: dsp.support@nlog.com Or visit
More informationMath 142, Exam 1 Information.
Mth 14, Exm 1 Informtion. 9/14/10, LC 41, 9:30-10:45. Exm 1 will be bsed on: Sections 7.1-7.5. The corresponding ssigned homework problems (see http://www.mth.sc.edu/ boyln/sccourses/14f10/14.html) At
More informationSmall Business Networking
Why network is n essentil productivity tool for ny smll business Effective technology is essentil for smll businesses looking to increse the productivity of their people nd processes. Introducing technology
More informationLecture 7: Integration Techniques
Lecture 7: Integrtion Techniques Antiderivtives nd Indefinite Integrls. In differentil clculus, we were interested in the derivtive of given rel-vlued function, whether it ws lgeric, eponentil or logrithmic.
More informationPresentation Martin Randers
Presenttion Mrtin Rnders Outline Introduction Algorithms Implementtion nd experiments Memory consumption Summry Introduction Introduction Evolution of species cn e modelled in trees Trees consist of nodes
More informationIntroduction to Integration
Introduction to Integrtion Definite integrls of piecewise constnt functions A constnt function is function of the form Integrtion is two things t the sme time: A form of summtion. The opposite of differentition.
More informationAn Efficient Divide and Conquer Algorithm for Exact Hazard Free Logic Minimization
An Efficient Divide nd Conquer Algorithm for Exct Hzrd Free Logic Minimiztion J.W.J.M. Rutten, M.R.C.M. Berkelr, C.A.J. vn Eijk, M.A.J. Kolsteren Eindhoven University of Technology Informtion nd Communiction
More informationUNIT 11. Query Optimization
UNIT Query Optimiztion Contents Introduction to Query Optimiztion 2 The Optimiztion Process: An Overview 3 Optimiztion in System R 4 Optimiztion in INGRES 5 Implementing the Join Opertors Wei-Png Yng,
More informationLECT-10, S-1 FP2P08, Javed I.
A Course on Foundtions of Peer-to-Peer Systems & Applictions LECT-10, S-1 CS /799 Foundtion of Peer-to-Peer Applictions & Systems Kent Stte University Dept. of Computer Science www.cs.kent.edu/~jved/clss-p2p08
More informationIf f(x, y) is a surface that lies above r(t), we can think about the area between the surface and the curve.
Line Integrls The ide of line integrl is very similr to tht of single integrls. If the function f(x) is bove the x-xis on the intervl [, b], then the integrl of f(x) over [, b] is the re under f over the
More informationThe Distributed Data Access Schemes in Lambda Grid Networks
The Distributed Dt Access Schemes in Lmbd Grid Networks Ryot Usui, Hiroyuki Miygi, Yutk Arkw, Storu Okmoto, nd Noki Ymnk Grdute School of Science for Open nd Environmentl Systems, Keio University, Jpn
More informationExam #1 for Computer Simulation Spring 2005
Exm # for Computer Simultion Spring 005 >>> SOLUTION
More informationStep-Voltage Regulator Model Test System
IEEE PES GENERAL MEETING, JULY 5 Step-Voltge Regultor Model Test System Md Rejwnur Rshid Mojumdr, Pblo Arboley, Senior Member, IEEE nd Cristin González-Morán, Member, IEEE Abstrct In this pper, 4-node
More informationMATH 2530: WORKSHEET 7. x 2 y dz dy dx =
MATH 253: WORKSHT 7 () Wrm-up: () Review: polr coordintes, integrls involving polr coordintes, triple Riemnn sums, triple integrls, the pplictions of triple integrls (especilly to volume), nd cylindricl
More informationsuch that the S i cover S, or equivalently S
MATH 55 Triple Integrls Fll 16 1. Definition Given solid in spce, prtition of consists of finite set of solis = { 1,, n } such tht the i cover, or equivlently n i. Furthermore, for ech i, intersects i
More informationIntegration. September 28, 2017
Integrtion September 8, 7 Introduction We hve lerned in previous chpter on how to do the differentition. It is conventionl in mthemtics tht we re supposed to lern bout the integrtion s well. As you my
More informationAlgorithm Design (5) Text Search
Algorithm Design (5) Text Serch Tkshi Chikym School of Engineering The University of Tokyo Text Serch Find sustring tht mtches the given key string in text dt of lrge mount Key string: chr x[m] Text Dt:
More informationAllocator Basics. Dynamic Memory Allocation in the Heap (malloc and free) Allocator Goals: malloc/free. Internal Fragmentation
Alloctor Bsics Dynmic Memory Alloction in the Hep (mlloc nd free) Pges too corse-grined for llocting individul objects. Insted: flexible-sized, word-ligned blocks. Allocted block (4 words) Free block (3
More informationComputing offsets of freeform curves using quadratic trigonometric splines
Computing offsets of freeform curves using qudrtic trigonometric splines JIULONG GU, JAE-DEUK YUN, YOONG-HO JUNG*, TAE-GYEONG KIM,JEONG-WOON LEE, BONG-JUN KIM School of Mechnicl Engineering Pusn Ntionl
More informationCOMP 423 lecture 11 Jan. 28, 2008
COMP 423 lecture 11 Jn. 28, 2008 Up to now, we hve looked t how some symols in n lphet occur more frequently thn others nd how we cn sve its y using code such tht the codewords for more frequently occuring
More informationSOME EXAMPLES OF SUBDIVISION OF SMALL CATEGORIES
SOME EXAMPLES OF SUBDIVISION OF SMALL CATEGORIES MARCELLO DELGADO Abstrct. The purpose of this pper is to build up the bsic conceptul frmework nd underlying motivtions tht will llow us to understnd ctegoricl
More informationPerformance analysis of QoS mechanisms in IP networks
University of Wollongong Reserch Online Fculty of Informtics - Ppers (Archive) Fculty of Engineering nd Informtion Sciences 2000 Performnce nlysis of QoS mechnisms in IP networks D. Ji University of Wollongong
More informationA TRIANGULAR FINITE ELEMENT FOR PLANE ELASTICITY WITH IN- PLANE ROTATION Dr. Attia Mousa 1 and Eng. Salah M. Tayeh 2
A TRIANGLAR FINITE ELEMENT FOR PLANE ELASTICITY WITH IN- PLANE ROTATION Dr. Atti Mous nd Eng. Slh M. Teh ABSTRACT In the present pper the strin-bsed pproch is pplied to develop new tringulr finite element
More informationMA1008. Calculus and Linear Algebra for Engineers. Course Notes for Section B. Stephen Wills. Department of Mathematics. University College Cork
MA1008 Clculus nd Liner Algebr for Engineers Course Notes for Section B Stephen Wills Deprtment of Mthemtics University College Cork s.wills@ucc.ie http://euclid.ucc.ie/pges/stff/wills/teching/m1008/ma1008.html
More informationComplete Coverage Path Planning of Mobile Robot Based on Dynamic Programming Algorithm Peng Zhou, Zhong-min Wang, Zhen-nan Li, Yang Li
2nd Interntionl Conference on Electronic & Mechnicl Engineering nd Informtion Technology (EMEIT-212) Complete Coverge Pth Plnning of Mobile Robot Bsed on Dynmic Progrmming Algorithm Peng Zhou, Zhong-min
More informationSection 10.4 Hyperbolas
66 Section 10.4 Hyperbols Objective : Definition of hyperbol & hyperbols centered t (0, 0). The third type of conic we will study is the hyperbol. It is defined in the sme mnner tht we defined the prbol
More informationCHANGING STRATA AND SELECTION PROBABILITIES* Leslie Kish, The University of Michigan. Summary
124 CHANGING STRATA AND SELECTION PROBABILITIES* Leslie Kish, The University of Michign Summry Survey smples re often bsed on primry units selected from initil strt with probbilities proportionl to initil
More informationMath 464 Fall 2012 Notes on Marginal and Conditional Densities October 18, 2012
Mth 464 Fll 2012 Notes on Mrginl nd Conditionl Densities klin@mth.rizon.edu October 18, 2012 Mrginl densities. Suppose you hve 3 continuous rndom vribles X, Y, nd Z, with joint density f(x,y,z. The mrginl
More informationDynamic Programming. Andreas Klappenecker. [partially based on slides by Prof. Welch] Monday, September 24, 2012
Dynmic Progrmming Andres Klppenecker [prtilly bsed on slides by Prof. Welch] 1 Dynmic Progrmming Optiml substructure An optiml solution to the problem contins within it optiml solutions to subproblems.
More informationOn the Detection of Step Edges in Algorithms Based on Gradient Vector Analysis
On the Detection of Step Edges in Algorithms Bsed on Grdient Vector Anlysis A. Lrr6, E. Montseny Computer Engineering Dept. Universitt Rovir i Virgili Crreter de Slou sin 43006 Trrgon, Spin Emil: lrre@etse.urv.es
More informationBefore We Begin. Introduction to Spatial Domain Filtering. Introduction to Digital Image Processing. Overview (1): Administrative Details (1):
Overview (): Before We Begin Administrtive detils Review some questions to consider Winter 2006 Imge Enhncement in the Sptil Domin: Bsics of Sptil Filtering, Smoothing Sptil Filters, Order Sttistics Filters
More informationIntroduction. Chapter 4: Complex Integration. Introduction (Cont d)
Introduction Chpter 4: Complex Integrtion Li, Yongzho Stte Key Lbortory of Integrted Services Networks, Xidin University October 10, 2010 The two-dimensionl nture of the complex plne required us to generlize
More informationDetermining Single Connectivity in Directed Graphs
Determining Single Connectivity in Directed Grphs Adm L. Buchsbum 1 Mrtin C. Crlisle 2 Reserch Report CS-TR-390-92 September 1992 Abstrct In this pper, we consider the problem of determining whether or
More informationWhat do all those bits mean now? Number Systems and Arithmetic. Introduction to Binary Numbers. Questions About Numbers
Wht do ll those bits men now? bits (...) Number Systems nd Arithmetic or Computers go to elementry school instruction R-formt I-formt... integer dt number text chrs... floting point signed unsigned single
More informationWhat are suffix trees?
Suffix Trees 1 Wht re suffix trees? Allow lgorithm designers to store very lrge mount of informtion out strings while still keeping within liner spce Allow users to serch for new strings in the originl
More information9.1 apply the distance and midpoint formulas
9.1 pply the distnce nd midpoint formuls DISTANCE FORMULA MIDPOINT FORMULA To find the midpoint between two points x, y nd x y 1 1,, we Exmple 1: Find the distnce between the two points. Then, find the
More informationApproximate computations
Living with floting-point numers Stndrd normlized representtion (sign + frction + exponent): Approximte computtions Rnges of vlues: Representtions for:, +, +0, 0, NN (not numer) Jordi Cortdell Deprtment
More informationTree Structured Symmetrical Systems of Linear Equations and their Graphical Solution
Proceedings of the World Congress on Engineering nd Computer Science 4 Vol I WCECS 4, -4 October, 4, Sn Frncisco, USA Tree Structured Symmetricl Systems of Liner Equtions nd their Grphicl Solution Jime
More informationEngineer To Engineer Note
Engineer To Engineer Note EE-186 Technicl Notes on using Anlog Devices' DSP components nd development tools Contct our technicl support by phone: (800) ANALOG-D or e-mil: dsp.support@nlog.com Or visit
More informationQuestions About Numbers. Number Systems and Arithmetic. Introduction to Binary Numbers. Negative Numbers?
Questions About Numbers Number Systems nd Arithmetic or Computers go to elementry school How do you represent negtive numbers? frctions? relly lrge numbers? relly smll numbers? How do you do rithmetic?
More informationInternational Journal of Scientific & Engineering Research, Volume 4, Issue 12, December ISSN
Interntionl Journl of Scientific & Engineering Reserch, Volume 4, Issue 1, December-1 ISSN 9-18 Generlised Gussin Qudrture over Sphere K. T. Shivrm Abstrct This pper presents Generlised Gussin qudrture
More information)
Chpter Five /SOLUTIONS Since the speed ws between nd mph during this five minute period, the fuel efficienc during this period is between 5 mpg nd 8 mpg. So the fuel used during this period is between
More informationCompanion Mathematica Notebook for "What is The 'Equal Weight View'?"
Compnion Mthemtic Notebook for "Wht is The 'Equl Weight View'?" Dvid Jehle & Brnden Fitelson July 9 The methods used in this notebook re specil cses of more generl decision procedure
More informationIn the last lecture, we discussed how valid tokens may be specified by regular expressions.
LECTURE 5 Scnning SYNTAX ANALYSIS We know from our previous lectures tht the process of verifying the syntx of the progrm is performed in two stges: Scnning: Identifying nd verifying tokens in progrm.
More informationCone Cluster Labeling for Support Vector Clustering
Cone Cluster Lbeling for Support Vector Clustering Sei-Hyung Lee Deprtment of Computer Science University of Msschusetts Lowell MA 1854, U.S.A. slee@cs.uml.edu Kren M. Dniels Deprtment of Computer Science
More informationChapter Spline Method of Interpolation More Examples Electrical Engineering
Chpter. Spline Method of Interpoltion More Exmples Electricl Engineering Exmple Thermistors re used to mesure the temperture of bodies. Thermistors re bsed on mterils chnge in resistnce with temperture.
More informationIntegration. October 25, 2016
Integrtion October 5, 6 Introduction We hve lerned in previous chpter on how to do the differentition. It is conventionl in mthemtics tht we re supposed to lern bout the integrtion s well. As you my hve
More informationFall 2018 Midterm 2 November 15, 2018
Nme: 15-112 Fll 2018 Midterm 2 November 15, 2018 Andrew ID: Recittion Section: ˆ You my not use ny books, notes, extr pper, or electronic devices during this exm. There should be nothing on your desk or
More informationRevisiting the notion of Origin-Destination Traffic Matrix of the Hosts that are attached to a Switched Local Area Network
Interntionl Journl of Distributed nd Prllel Systems (IJDPS) Vol., No.6, November 0 Revisiting the notion of Origin-Destintion Trffic Mtrix of the Hosts tht re ttched to Switched Locl Are Network Mondy
More informationFall 2018 Midterm 1 October 11, ˆ You may not ask questions about the exam except for language clarifications.
15-112 Fll 2018 Midterm 1 October 11, 2018 Nme: Andrew ID: Recittion Section: ˆ You my not use ny books, notes, extr pper, or electronic devices during this exm. There should be nothing on your desk or
More informationRepresentation of Numbers. Number Representation. Representation of Numbers. 32-bit Unsigned Integers 3/24/2014. Fixed point Integer Representation
Representtion of Numbers Number Representtion Computer represent ll numbers, other thn integers nd some frctions with imprecision. Numbers re stored in some pproximtion which cn be represented by fixed
More informationOverview. Network characteristics. Network architecture. Data dissemination. Network characteristics (cont d) Mobile computing and databases
Overview Mobile computing nd dtbses Generl issues in mobile dt mngement Dt dissemintion Dt consistency Loction dependent queries Interfces Detils of brodcst disks thlis klfigopoulos Network rchitecture
More informationINTRODUCTION TO SIMPLICIAL COMPLEXES
INTRODUCTION TO SIMPLICIAL COMPLEXES CASEY KELLEHER AND ALESSANDRA PANTANO 0.1. Introduction. In this ctivity set we re going to introduce notion from Algebric Topology clled simplicil homology. The min
More informationAccelerating 3D convolution using streaming architectures on FPGAs
Accelerting 3D convolution using streming rchitectures on FPGAs Hohun Fu, Robert G. Clpp, Oskr Mencer, nd Oliver Pell ABSTRACT We investigte FPGA rchitectures for ccelerting pplictions whose dominnt cost
More informationUnit 5 Vocabulary. A function is a special relationship where each input has a single output.
MODULE 3 Terms Definition Picture/Exmple/Nottion 1 Function Nottion Function nottion is n efficient nd effective wy to write functions of ll types. This nottion llows you to identify the input vlue with
More informationData sharing in OpenMP
Dt shring in OpenMP Polo Burgio polo.burgio@unimore.it Outline Expressing prllelism Understnding prllel threds Memory Dt mngement Dt cluses Synchroniztion Brriers, locks, criticl sections Work prtitioning
More informationPARALLEL AND DISTRIBUTED COMPUTING
PARALLEL AND DISTRIBUTED COMPUTING 2009/2010 1 st Semester Teste Jnury 9, 2010 Durtion: 2h00 - No extr mteril llowed. This includes notes, scrtch pper, clcultor, etc. - Give your nswers in the ville spce
More informationScalable Distributed Data Structures: A Survey Λ
Sclble Distributed Dt Structures: A Survey Λ ADRIANO DI PASQUALE University of L Aquil, Itly ENRICO NARDELLI University of L Aquil nd Istituto di Anlisi dei Sistemi ed Informtic, Itly Abstrct This pper
More informationDigital Design. Chapter 6: Optimizations and Tradeoffs
Digitl Design Chpter 6: Optimiztions nd Trdeoffs Slides to ccompny the tetbook Digitl Design, with RTL Design, VHDL, nd Verilog, 2nd Edition, by Frnk Vhid, John Wiley nd Sons Publishers, 2. http://www.ddvhid.com
More informationScanner Termination. Multi Character Lookahead. to its physical end. Most parsers require an end of file token. Lex and Jlex automatically create an
Scnner Termintion A scnner reds input chrcters nd prtitions them into tokens. Wht hppens when the end of the input file is reched? It my be useful to crete n Eof pseudo-chrcter when this occurs. In Jv,
More informationNearest Keyword Set Search in Multi-dimensional Datasets
Nerest Keyword Set Serch in Multi-dimensionl Dtsets Vishwkrm Singh Deprtment of Computer Science University of Cliforni Snt Brbr, USA Emil: vsingh014@gmil.com Ambuj K. Singh Deprtment of Computer Science
More informationConvergence of Wachspress coordinates: from polygons to curved domains
Convergence of Wchspress coordintes: from polygons to curved domins Jiří Kosink Computer Lbortory University of Cmbridge Jiri.Kosink@cl.cm.c.uk Michel Brtoň Numericl Porous Medi Center King Abdullh University
More informationA Progressive Register Allocator for Irregular Architectures
A Progressive Register Alloctor for Irulr Architectures Dvid Koes nd Seth Copen Goldstein Computer Science Deprtment Crnegie Mellon University {dkoes,seth}@cs.cmu.edu Abstrct Register lloction is one of
More informationMisrepresentation of Preferences
Misrepresenttion of Preferences Gicomo Bonnno Deprtment of Economics, University of Cliforni, Dvis, USA gfbonnno@ucdvis.edu Socil choice functions Arrow s theorem sys tht it is not possible to extrct from
More informationCS : Data Structures Michael Schatz. Nov 7, 2018 Lecture 29: HashTables pt2
CS 600.226: Dt Structures Michel Schtz Nov 7, 2018 Lecture 29: HshTbles pt2 HW7 Assignment 7: Whispering Trees Out on: November 2, 2018 Due by: November 9, 2018 before 10:00 pm Collbortion: None Grding:
More informationCSCI 104. Rafael Ferreira da Silva. Slides adapted from: Mark Redekopp and David Kempe
CSCI 0 fel Ferreir d Silv rfsilv@isi.edu Slides dpted from: Mrk edekopp nd Dvid Kempe LOG STUCTUED MEGE TEES Series Summtion eview Let n = + + + + k $ = #%& #. Wht is n? n = k+ - Wht is log () + log ()
More informationWhat do all those bits mean now? Number Systems and Arithmetic. Introduction to Binary Numbers. Questions About Numbers
Wht do ll those bits men now? bits (...) Number Systems nd Arithmetic or Computers go to elementry school instruction R-formt I-formt... integer dt number text chrs... floting point signed unsigned single
More informationImproper Integrals. October 4, 2017
Improper Integrls October 4, 7 Introduction We hve seen how to clculte definite integrl when the it is rel number. However, there re times when we re interested to compute the integrl sy for emple 3. Here
More informationThe Fundamental Theorem of Calculus
MATH 6 The Fundmentl Theorem of Clculus The Fundmentl Theorem of Clculus (FTC) gives method of finding the signed re etween the grph of f nd the x-xis on the intervl [, ]. The theorem is: FTC: If f is
More informationPhysics 152. Diffraction. Difrraction Gratings. Announcements. Friday, February 2, 2007
ics Fri Feb.02. Announcements Diffrction Difrrction Grtings Fridy, Februry 2, 2007 Help sessions: W 9-10 pm in NSC 118 Msteringics WU #5 due Mondy WU #6 due Wednesdy http://www.voltnet.com/ldder/ A bem
More informationOn Computation and Resource Management in Networked Embedded Systems
On Computtion nd Resource Mngement in Networed Embedded Systems Soheil Ghisi Krlene Nguyen Elheh Bozorgzdeh Mjid Srrfzdeh Computer Science Deprtment University of Cliforni, Los Angeles, CA 90095 soheil,
More informationII. THE ALGORITHM. A. Depth Map Processing
Lerning Plnr Geometric Scene Context Using Stereo Vision Pul G. Bumstrck, Bryn D. Brudevold, nd Pul D. Reynolds {pbumstrck,brynb,pulr2}@stnford.edu CS229 Finl Project Report December 15, 2006 Abstrct A
More informationStack. A list whose end points are pointed by top and bottom
4. Stck Stck A list whose end points re pointed by top nd bottom Insertion nd deletion tke plce t the top (cf: Wht is the difference between Stck nd Arry?) Bottom is constnt, but top grows nd shrinks!
More information