You are on page 1of 25

SortingAlgorithms

http://www.ccs.neu.edu/jpt/archive/2.6.0/applets/PlayingCards/applet.htm

AdamMeyers,NYU June,2008
MUSICOMPUTATION SortingAlgorithms June,2008

Outline
WhatisanAlgorithm? WhatisSorting? WhatisaSortingAlgorithm? SomeSortingAlgorithms AsymptoticComplexity Summary
MUSICOMPUTATION SortingAlgorithms June,2008

WhatisanAlgorithm?
Arecipeforsolvingaproblem Algorithmscanbeimplemented
Bydifferentprograms Indifferentprogramminglanguages

Differentalgorithmsforsolvingthesameproblem
Aremoreefficientiftheyrequirefewerbasicoperations Maysolvedifferentinstancesofthesameproblemmore orlessefficiently
MUSICOMPUTATION SortingAlgorithms June,2008

WhatisSorting?
Sorting=CreateasequenceofitemsinaSetS
Suchthatabinaryrelation(e.q.,)holdsforpairsof consecutiveelements

WeassumeCARD1<CARD2iff
FaceValue(CARD1)<FaceValue(CARD2)
2<3<4<...<J<Q<K<A

FaceValue(CARD1)=FaceValue(CARD2)AND SuitValue(CARD1)<SuitValue(CARD2)
CLUBS<DIAMONDS<HEARTS<SPADES
MUSICOMPUTATION SortingAlgorithms June,2008

WhatisaSortingAlgorithm?
Analgorithmforsortingasetofelements Differentalgorithmsmaybebetter(more efficient)forsortingunderdifferent circumstances

MUSICOMPUTATION SortingAlgorithms June,2008

InsertionSort
Startwithashuffleddeckofcards
Deck=Card1,Card2,...Card52

ChooseaspottostoreyourResult
Result=asequenceof0ormorecards Resultisinitiallyempty

ForeachCardNfromCard1toCard52
Followtheinstructionsonthenextslide

MUSICOMPUTATION SortingAlgorithms June,2008

TraversetheCardsinResultfromrighttoleft comparingCardMinResulttoCardN IfCardNhasahighervaluethanCardM


PutCardNtotherightofCardM(andstop)

InsertionSort:ForeachCardN

IfthecardsrunoutbeforeplacingCardN
InsertCardNatthefrontofResult

Otherwise,continuecomparingcards Intheworstcase,all52cardswillbecomparedto anaverageof51/2cards(thecardsintheResultat thetime)(52X51/2=1326comparisons)


MUSICOMPUTATION SortingAlgorithms June,2008

BubbleSort
DoBubbleuntilthecardsaresorted,i.e.,untilthe stepsdonotresultinanyswappingofcards. Bubble:ForeachCardNbetweenCard1..51
IfCARDN>CARDN+1,swaptheorderofthetwocards (CARDNbecomesCARDN+1)

ABubbleinvolves51/2comparisonsonaverage 51bubbleswilltakeyouthroughtheentirestack Acompletesorttakes51X51/2=1300.5 comparisonsintheworstcase


MUSICOMPUTATION SortingAlgorithms June,2008

MergeSort
Ifthedeckhasonlyonecard
Thereisnothingtodo

Otherwise
Dividethedeckofcardsinhalf MergeSorteachhalf Mergeeachhalftogether
Mergeisdefinedonthenextslide
MUSICOMPUTATION SortingAlgorithms June,2008

Merge(forusebyMergeSort)
Inputconsistsof2SortedPilesofCards:
Pile1andPile2

BeginwithanemptyResultanddothe followinguntiloneorbothpilesareempty
ComparethetopcardsofPile1andPile2and placethelowestcardfacedowninResult

IfoneofthetwoPilesisnonempty,placethe nonemptypileattheendoftheResult
MUSICOMPUTATION SortingAlgorithms June,2008

MergeSort:NumberofComparisons
Foranysetofnthings,youcandividethosen thingsinhalflog2ntimesbeforeyouendupwith just1thingineachhalf(approximately) Eachsplitcorrespondstomergesinvolving51 comparisons(seeMergeinthepreviousslide) Thereareapproximately51Xlog252291 comparisons
MUSICOMPUTATION SortingAlgorithms June,2008

Ifthedeckhasonlyonecard,donothing Ifthedeckhastwocards,puttheminorder Otherwise


ChooseacardandsetitasthePIVOTandcreateinitially emptystacksLEFTandRIGHT ForeachremainingCARDNinthedeck
IfCARDNprecedesPIVOT,placeCARDNinLEFT ELSE,placeCARDNinRIGHT

QuickSort

QuickSortLEFTandRIGHT Formasinglestack
LEFTontopofPIVOTontopofRIGHT
MUSICOMPUTATION SortingAlgorithms June,2008

QuickSort2
QuickSortinvolvesintheballparkof52 log252=296comparisonsonaverage QuickSortinvolves52X52=2704 comparisonsintheworstcase. Typicallyittendstohavefewercomparions thanMergeSort(andotheralgorithms)and isthusassumedtobequicker

MUSICOMPUTATION SortingAlgorithms June,2008

SimilartoBucketSortandCountingSort Mark13positions(buckets)infrontofyou Gothroughall52cardsand


Puttheclubsinthefirstbucket Putthediamondsinthesecondbucket Puttheheartsinthethirdbucket Putthespadesinthefourthbucket

RadixSort

Combinethestacks:
firstontopofsecondontopofthirdontopoffourth

NextSlide

MUSICOMPUTATION SortingAlgorithms June,2008

Gothroughall52cardsand

RadixSort2

Putthe2sinthefirstbucket Putthe3sinthesecondbucket ... PuttheAcesinthe13thbucket

Combinethestacks:
firstontopofsecondontopofthird,...,12thontopof13th

I.e.,sorttheleastsignificantdigitorequivalentfirst
MUSICOMPUTATION SortingAlgorithms June,2008

Algorithmsarejudgedby

Complexity

Time(howfast) Space(howmuchmemory)

Asymptoticcomplexity
Asdatasizeincreases,howfastdoestime/spaceuse increase

UpperBoundcomplexity(BigO)
Worsecase

Othercases:
LowerBound Average
MUSICOMPUTATION SortingAlgorithms June,2008

AsymptoticTimeComplexity
Countnumberofbasicoperations
Comparisonsoftwocards Insertionsofcards

Evaluatingsortingalgorithms
Asthesizeofthedeckofcardsincreases,
Howmanymorecomparisonsandinsertionsarethere?

MUSICOMPUTATION SortingAlgorithms June,2008

ComplexityisApproximate
Whenyouweighanelephant,doyouconsiderthe weightofaflyonitsback? Complexity
Expressedasapolynominal BigOnotationtakeshighestterm,ignoringconstants 4X3+X23becomesO(X3)

MUSICOMPUTATION SortingAlgorithms June,2008

FiguringoutComplexity
InMergeSort(supposen=52)
yousplitthedecklog2ntimes
Fora52carddeck,youhavetodoabout6splitsbefore youhavejust1cardinapile

Thenforeachsplitofthewholedeck,youmustdo n1comparisons(let'sroundoffton)
i.e.thereare51comparisonsfora52carddeck

ncomparisonsXlog2nsplitsmeansthat... MergeSorthasannlog2ntimecomplexity
Approximately300comparisonsfor52cards
MUSICOMPUTATION SortingAlgorithms June,2008

PatternofGrowth
n 1 2 3 4 5 6 logn 0 1 1.6 2 2.3 2.6 2n 2 4 6 8 10 12 nlogn 0 2 4.8 8 11.6 15.5 N2 1 4 9 16 25 36 2n 2 4 8 16 32 64

MUSICOMPUTATION SortingAlgorithms June,2008

GraphingComplexities

MUSICOMPUTATION SortingAlgorithms June,2008

AsymptoticComplexityisanEstimate
ActualComlexity Asymptotic1 Asymptotic2 3n7000 3n exponential n3+n22 n3 cubic 5n2+3n+7 n2 quadratic n2+(nlogn) n2 quadratic 5n+logn3 5n linear 3logn50 3(logn) logarithmic
MUSICOMPUTATION SortingAlgorithms June,2008

Complexities
LogarithmicTime:Constant*logn LinearTime:Constant*n nlognTime PolynomialTime:nmwheremisaconstant
n2orn3orn4... quadratic,cubic,etc.

ExponentionalTime:m wheremisaconstant
2nor3nor4n..
MUSICOMPUTATION SortingAlgorithms June,2008

InsertionSort,BubbleSort:O(n2) MergeSort:O(nlogn) QuickSort:O(nlogn)


ifimplementedbadlyO(n2)
Choosingbadpivot

SortingAlgorithmComplexities

TypicallyfasterthanMergeSort

RadixSort:O(kn),wherek=numberofbins
lineartimealgorithm useslotsofspace requiresdetailedknowledgeofinput
MUSICOMPUTATION SortingAlgorithms June,2008

UsefulLinks
Descriptions en.wikipedia.org/wiki/Sort_algorithm
www.personal.kent.edu/~rmuhamma/Algorithms/algorithm.htm

AnimatedDepictionsandApplets
vision.bc.edu/~dmartin/teaching/sorting/animhtml/all.html

homepage.mac.com/mihailod/atic/sorting.html coderaptors.com/?All_sorting_algorithms

MUSICOMPUTATION SortingAlgorithms June,2008

You might also like