Graphs and Sorting SYLLABUS Graphs: Graph Implementation Methods, Graph Traversal Methods. Serting: Quick Sort, Heep Sort, External Sorting — Medel for External Sorting, Merge Sort. LEARNING OBJECTIVES Graph and its Types Bosic Terminology Related to Graphs Vorious Operations Performed on Graphs Various Types of Graph Traversal Methods . Quick Sort Technique Heap Sorting Technique * Process of External Sorting External Sorting Mode! SOR RURAL AES Merge Sort and its Algorithm. INTRODUCTION es) and set of ares (edges). Every V, E) where, V ‘A graph is @ data structure that consists of set of nodes (ver edge present in the graph is indicated by a pair of vertices. Itis defined as G ‘a the set of elements called nodes or vertices or points. And Eis the set of edges of the graph identined with a unique pair (U, V.) of nodes. The various operations that can be performed ‘on graphs are Insertion, Deletion, Traversing and Merging. Sorting is 0 technique of organizing the data or arranging the records either in ascending or descending order (Le, bringing soe orderliness In the data), Sorting can be performed on any one or combination of ene or more aitributes present in ‘each record. It is very easy and efficient to perform searching, if data Is stored in sorted order It's classified into internal and ‘external sorting, Heap sort fs also known as selection sort algorithm which is on internal sorting technique, And Merge sort is an external sorting techniqve which makes use of secondary, storage. SPECTRUM ALL-IN-ONE JOURNAL FOR EN' IEERING STUDENTS420 DATA STRUCTURES [JNT| 'U-HYDERABAD) 'Q2. Define graph. DI juss thé differant types of graphs. ; oR What is Graph? Define degree of vertox, Doc.-19(R 18), a1() Q1. Explain in detail about graph ADT. Answer = Answer: Graph Graph ADT 'A graph is defined as G = (V. E) where, ‘A craph is @ data structure that consists of set | (i) is the set of elements called nodes or vertices or points. of nodes (vertices) and set of ares (edges). Every edge present in the graph is indicated by a pair of vertices. ‘The various graph ADT operations are as follows, ages of the graph identified with V) of nodes. cotes that there is an edge Eis the set of € ‘a unique pair (U, Here (U,, ¥) pair den 1. Create(): This method is used to create an empty Guph The created graph does not contain any | fromnode Ute node vertices and edges. Different Types of Graphs , aes eneatteaph, vel): This method is used | (a) Directed Graph: A directed ‘graph is a graph in to insert 2 new vertex vel into the graph. The eve the pair of vertices that make up an edge inserted vertex does not contain’any adjacent wre ordered, In such graph, the order of vertices edges. representing an edge is important. 3. DeleteVertex(graph, vel): This method is used _ Example to delete an existing vertex vel from the graph. ‘All the adjacent edges that are connected with @ ©) vertex vel are also being deleted. 4. InsertEdge(graph, ve2, ve3): This method incerts an edge e! between the vertices ve2 and ve3 5, DeleteEdge(graph, ve2, ve3): This method @) (c) deletes an edge e] existing between the vertices vve2 and ve3. Figure: Directed Graph 6 Boolean IsEmpty(graph): This method checks , Here, (4, B) and (B, A) represents two distinc whether the graph is empty or not. If empty, edges. returns true, otherwise returns false. (>) Undirected Graph: In an undirected graph, the 7, List Adjacent(graph, vel): This method returns order of pair of vertices is not important. ail the respective edges that are adjacent to the ster ©) ‘The program for implementing graph ADT is as follows, class GraphADT { public: . @) (c) a ‘GraphADT() Figure: Undirected Graph Here (4, B) and (B, A) represent the same edge (© Mixed Graph: A graph in which some edges ®™ ‘WARNING: Xerox/Pht 3 bool isEmpty( ) Const {return vertices = 0}; int NumberOfVertices() Const {return vertices}; int NumberOfEdges( ) Const {return edge} virtual int Degree (int p) Const = 0; virtual bool existEdge (int p, int q) Const = virtual void insert Vertex (int q) virtual void insertEdge (int p, intq) virtual void deleteVertex (int q) = 0; virtual void deleteEdge (int p, int ) private: int vertices; int edges; jotocopying of this book is. ‘directed and some edges are undirected is know? asa mixed graph. Examplewitea: Graphs and Sorting Ta(V. £) be a graph and let x € E be a directed associated with thé pair of nodes (U,V). The edge ie eid to be initiating or originating in node U and ‘eeminating ‘or ending in node V. The nodes Uand V are tem alled initial node and terminal nodes of edge ose an edge of @ graph which connects t called a loop. (y~ loop Example edge x Inthe above graph, for the edge x, 4 is the initial node and B is the terminating node. Degree of Vertex Degree of vertex (v’) in Graph G is defined as the number of edges incident on vertex v, self-loop counts two edges. Consider an example, S_@Q_* elf is Figure In the above figure, Degree of vertex v, is 4 Degree of vertex v, is 3. Explain about connected and no: connected graph and list the differences between them. Answer : Connected and Non-connected Graphs Ina graph, ifthere exist a path between every pair, of vertices then the graph is known as connected graph. Inaconnected graph, it is possible to traverse from one node to another node. On the other hand, ifno path exist between any pair of vertices then the graph is known as non-connected graph. Example Q3. SPECTROM ALL-IN-ONE JOU! IRNAL FOR ENGINEERING STUDENTS il The graph is connected because thete from each vertex to every other vertex. Path A to B:A +> B AtoC:A+B DC. AtoD:A+~B>D O—®) Figure: Non-connected Graph “The graph is non-connected because there is no path from A to D, there is no path from C to any other node. Differences between Connected and Non-connected Graphs ‘A directed graph is said to be strongly connected if there exist a path from every vertex to every other vertex. On the other hand, a directed graph is said to be weakly connected if two or more vertices in the graph are not connected. ‘An undirected graph is called a connected graph if every node in the graph can be reached from any other node. Graphically, a connected undirected graph consist of single ‘connected component which is a connected sub graph. Q4. Write short notes on the following graph terminology, (a) Subgraph —(b) Path (c) Tree (d) Cycle (e) Parallel edges (f). Complete graph (g) Cyclic and Acyclic graph. Answer : (@) Subgraph Consider two graphs G and G,, say G, is subgraph of G if, (@ All the vertices and all the edges of G, are in G. (i) Each edge of G, has the same end vertices in G as in G,. A subgraph is a graph which is a part of another graph. vy <<). IN\ ve ee, Figure Itcan be observed that, all the vertices and edges of graph G, are in graph G and also that every edge in G, has the same end vertices in G as in G,. So, it can be concluded that G, is a subgraph of G.112 DATA STRUCTURES. [UNTU-HYDERABAD) () Path A path can be defined as a trail in which no vertex appears more than once. In other words ifall the vertices ina trail are distinct, then the trail is called a path. If in an open walk no edge appears more than once, then the walk is called a trail. (©) Tree A tree is defined as a finite set of one or more ‘elements with one element designated as root and the other elements (if exists) are divided into trees are called subtrees. Figure belowillustrates some ofthe examples of tres. Q ® Q oO , @ ©—O oO © ©O—® © Figure: Examples of Trees @ Cycle . Itcan be defined as a circuit in which the terminal vertex does not, appear as an internal vertex and no + internal vertex is repeated. A circuit is a closed walk, where no edge appears more than once. (©) Parallel Edges Ifa pair of vertices contains more than one edge then the edges are called as parallel edges. The graph will be called as multigraph in such cases. (© Complete Graph ; Ifa vertex contains edges to al the vertices froma it, then the graph is called complete graph. (@), Cyclic and Acyelic Graph Ifa path contains edges starting from a vertex and ending at the same vertex, then this path is called «as eycle. The greph will be called as cyclic graph. If'a ‘eraph does not contain any eycles, then such graph will be called as acyclic graph. 5. Write short notes on the following repre- sentations, (a) Adjacency matrix (b) Adjacency list (c) Adjacency multilists. OR How to represent graphs? Explain. Dec.-19(R18), 08 oR ive example for adjacency list of a graph. (RE OR ToplesAmacerigy LD) Answer ‘AprilMay-23(R18), a1(h) (a) Adjacency Matrix ‘Anadjacency matrix A = (a,) of graph Gis defined a graph, “V, is adjacent to y> Ti case of directed graph “Yi i means that there isa directed ede from V7, 10 ¥, Example! Consider the undirected graph given below, ‘Adjacency matrix is, 1 2 3 4 1fpty' to A=2 [1 o | 0 i sfifo}o}3 stati to Example 2 : ‘Consider the directed graph shown below, Here a, has 1 when there is an edge from ¥,10, The adjacency matrix A shows the number of paths of length J between any pair of vertices (Y, V). The adjacency matrix 4? shows the number of paths of length 2 between any pair of vertices. Example 3 ___ Consider the adjacency matrix for the graph given in example 2. 1203 4 1{o; 1] oT} 0 4=2/o0[ofo] 1 3fa[ofo]1| 4loto/]ito ofifojo of1fofe ofofof{i] [ofolol! Ljofoji 1{oj{ol! ofofifo] fofofife 1234 1fofofo]1 oe? offi [o 3fofififo 4t1[ofof[1 WARNING:eaves that there is Tone path of length 2 from ma Hy Hts Ty to Vy Fy to Fyand ¥, to Fy Ren wqibe veritied from graph of example 2, ws & represents an adjacency matrix an generals 4 Remar OF aus ‘of length k present between ajacency Matrix Representation aback of At repent pap of nodes adjaceney matrix, age n> missed: For mast of the graphs, only few pe ited in mates, So, it ead fo memory For example, waste: ceney matrix for this graph is 1 4 1 0 0 0 0 0 0 0 0 3 0 1 0 1 afi jo 0 Here the size of matrix is 4 * 4= 16, but only seauies are filled with 1, Hence, a lot of memory wased, () Adjacency List In adjacency list repres =(V, B), each vertex of ‘graph is represented by 2° adjacency list. The adjacency list corresponding to saree contains all vertices that are adjacent © “vin | ‘G Inadjacency list: representation, each adjacency list Feared aca chain soit is called linked adjacency ist representation. ry ‘A’ is used such entation of a graph, G wa wi CoP wa ChE ES ™ CHE WA NOL Figure: Adjcency List Representation for Graph (G,) Figure: Directed Graph (6) wa CELE wa CoP wi w+ [pout] va (+ wi ee Figure: Adjacency List Repre sntation for Graph (G,) (©. Adjaceney Multilists ‘Adjaceney multilists are the lists in which nodes may be shared among multiple lists. In adjacency multilists, for every edge there is. ‘exactly one node. This rhode is present on the adjacency lists for the Wo nodes ineident upon it. Consider the following graph, In this representation, an a"Tay ‘tat Aff points tothe linked list that jneludes all vertices ‘which a particular vertex (V)isconnected, Each node in the linked list contains two com , vertex ed link, strates linked adjacency » The following figuresil ; list representations for directed rand undirected gr=phss fis, AY - ponents i.e. n ‘The adjacency multlists for this graph are as follows, TBR) sven , aden: 81 5 Grafs] 12 Noes ote ase: N5 Tali} na Notee na 98 Nisee N84 7] Tse ss GraphiG,) Figure: Undi RS SPECTRUM ALL-IN-ONE JOURNAL FOR ENGINEERING STUDENTSYDERABAD} explall at DATA STRUCTURES INT G7. What aro tho various oporations that can The format of node structure in adjacency ists is Link | Node 1 | Node 2 Link 2 ] Where node! and node? represents the shared edge. Link 1 and link 2 represents the next unvisited edges from node 1 and node 2. For example, nodes for shared edge NI are a and b, The next unvisited edges for those nodes are NS and Q6. Discuss the pros and cons of various graph representations. Answer > Graph Representations faph representations are as ‘The three common gt follows, 1. Adjacency Matrices Pros: The pros. or advantages of adjacency matrices are as follows, It is preferred for the graph storage where there exist less number of edges.» «+ Itefficiently determines the relations! ina graph. Con: For answer refer Unit-1V, Page No. 113, Q5, Topic: Drawback of Adjacency Matrix Representation. 2, Adjaceney Lists Pros: The pros or advantages of adjacency lists are as follows, ‘> It offers a compact representation for sparse graphs. 4 Ttean be directly used for weighted graphs. 4 Itoffers robustness of adopting various other repre- sentations. Cons: The cons or disadvantages of adjacency lists are as follows, 5 “It does not provide faster ways of determining whether a specific edge’exist in the graph or not. ‘ _Itiscomplex when compared to adjacency multil- ists 3. Adjacency Multilists Pros: The pros or advantages of adjacency multilists are as follows,” 4 Itallows sharing of nodes within multiple lists. + Itoffers simplified representation when compared _ toadjacency lists. Con The main con or disadvantage of adjacency multilist is that, it does not work well'when the ‘number of vertices are less. ; bo performed on graphs Model Papers, €9(b) Answer : “The vation mat can be performed are, \eva us operations th 1 Insertion a vertex into the graph fest allocate thes mory to the ne’ {then store the ut data rae ew Ver E eia data elements wrtex, Next, set all the met in the new vertex. Next, ee TONULL, While inserting the verte’ chetk whet te ie ertex must be inserted before viaph isempty or the new Ve be inserted befor thelist ‘vertex. In both the cases, the NEW vertex becomes the first node of the graph, If neither ‘of the ease is true then insert the new vertex in sequence, ‘To insert a ww vertex Algorithm: a —» data that has t sory for the new VErleX, /ERTEX*)malloc(size of VERTEX) Input: dat be inserted into vertex. Step1: Allocate met newvertex=(VI newvertex, Step2: Store datal newvertex — dataptr=data; Step3: Initialize meta data elements in new vertex, newvertex — outdegre newvertex —> indegree=0; newvertex —> processes newvertex — pArc=NULL; newvertex — ptrNextverte ‘Step: graph + count++; StepSeIf graph is empty, iffgraph — first NULL) Set the first vertex to new vertex, graph — first=newvertex; Step6:else ‘Search for the insertion point, prev=NULL; while(loc & (graph —> compare(data, loc —+ dataptr)>®)) pres 5 ‘ loc=loc —+ ptrNextvertex; Insertion vertex before the start node if(tprev) graph — start-newvertex;. else ‘ Insertion of a new vertex in sequence, pre— ptrNextvertex=newvertex; newvertex — ptrNextvertex=I Step7: end EE a cS SRT Tram ‘wi Tt acer ad eG amie this tava aaNetGraphs arid Sorting Deletion To delete a vertex, initially sea mgr itty er te ee ‘feck whether any are is ehtering (indegree) or leaving {outdegree) the vertex. A vertex with degree greater fan zero cannot be deleted. So to handle such a vertex, a check for indegree and outdegree in the vertex is ree quired. Algorithm Input: Key is the key of the vertex which has to be de- Ieted from a non empty graph, 7 stept:If graph is empty, Iffgraph — start =NULL) return-1; sstep2:Search for the location where key is present, prev-NUL pir=graph—first; while(ptr &&(graph — compare(key, ptr > dataptr)>0)), + prev=ptr;. pir=ptr — ptrNextvertex; ‘Step3:If vertex is not found, i ifttptr | graph > compare(key, ptr — dataptr)! ¢ a return -1;, Step4iIf vertex is found, test the degree, if ptr — indegree > 0 || ptr outdegree0) return ~2; “StepS:Delete the vertex, if(tprev) graph — first = ptr + ptrNextvertex; else prev — ptrNextvertex=ptr — pttNextvertex; Step6: Decrement the count after deletion, : graph — count; Step7:return 1 Step8:end 3. Traversing For answer refer Unit-IV, 4. Merging Merging operati graphs say G1 and G2 int Page No. 115, Q8. jon involves merging of two to a new graphs G3. In addition tb this, seme of the new entties are present With espa to new edges. So, merging can be done by copying adjacent matrices of the components and then setting new edges with entries. Algorithi “Stepl: Initialize the entries of the fori=1tondo forj=iton do DGpitlilfi) = 9 endFor endFor matrix of graph.G- to the adjacent 2 ‘Step2:Copy adjacent matrix of G1 matrix of G fori=1ton, do for j= 1 ton, do DGptfilfi) = DGIptLilLT endFor endFor Step3: Copy the adjacent matrix of G2 into the adjacent matrix of G for i=1 ton, do for j= 1 ton, do DGpt{i +n] + 4,) = DG2prli]ET endFor endFor : Step4:Sct the edges from G1 to G2 doFlag = TRUE While(doFlag) do Read(v, w) If{v #include ‘Header file section dinclude ##define MAX 20 typedef struct Queue { int data[MAX]; int nh; j Jqueue; typedef struct node ‘WARNING: KeroxIPhotocopying of this book iy a CRIMINAL act. Am 1 Number of nodes i int deque(queue *); int empty(queue *); void BFS(int); void graph(); /iCreate an adjacency list void insert(int v ¥2); ‘Mnsert an edge (v1,v2)in aK Yoid DFS(int i int vis[MAX]; node *G[20};, eteads of the linked list int fie BLE to faco LEGAL procesdi™®™graphs and Sorting we) Trin tani ius t ‘void DFSMina fy ait ‘ eet node * e Prinele“tad € iy, t rat; redial KCrede 2)BFS 3)DFS 4). visi} ~ gSelect your option: Krvie"y, whilelp!=NULL) option:"y, a Feed Lech, t gwachich) npovertex; { ifttvis{ify —— DSU bee: peponent; a ; case T: pref" Enter the starting node number“) 5 sealed Ki, aa : BES}. ‘void enque(queve *e,int x) Minar elements into the queve ban . i case 3: fortran) iffeon—=1) oooh e-mdatafern 4 else { concent; edataleon]=x; ; t ; ice wivislMAAL int deque(queue *c) queue g; ‘Delete elements from the queue quese G " rede *p; { int x garqhat; n me feniitrinii*) are ttalo ee if e>n—=c->h) equa), prin Md 0); visfv}=1; whildtenpty(Zq)) t cehec-rhtl; wodequcdéeay; return(x); forlp-G{v]p!“NULL sp P79 } ‘Menara urvisited abjaces vertices Of vio que void graph() t ( wep vertex; int il v2,¢; ittvisfu==0) printf(“Enter number of vertices 2") t seant("%d" den); for(i=Oii ‘MHleader file section Void insert(int v1,int v2) int q[20],t=- Lye { 1iGlobal declaration of variables node *p.*q: int A{10][10],B(20},Stk(20]; g=(node *)malloc(sizeof{node)); . 4 Acquire memory for the new node wold éreate(int item}: ae void BFS(int x,int y); qonext=NULL: void DFS(int x,int y): iRGIvI}=NULL) void push_item(int item); insert the . int pop_itemO; eect inthe linked list for the vertex.no. VI void main) ‘/imain Function ce { ao { int num,,s,choice; - char od; p=Glv i clrser(); while(p-Snex ULL) printf(*Pleas¢ enter the total number ofvertices.") _ Goro the end of linked list” scanfi"%a”, num); pep>next; num:i+) ponext=q; . ; e for(j=1;j<-num;j+) { int empty(queue *c) 5 printf(“Press 1 if Yd had a Node with %d else { press 0:",ij); ite>a—1) scanf(“%d",&ALILDs retum(1); } return(0); } } printf(“The Adjaceney Matrix is:\n”); for(i=l;i<-num;i++) { for(j=1;j<-num;j+1) { printf %d”,ALGDs ¢ eae ar ac Gime carats: printf(\n"); Seesmic } - do { ped for(i=: Peete Bli roe rae Ro, printf(‘\n MENU: 1.BFS 2.DFS” artemis e a printf(“\n Select your option:”); pone scanf(“%d"éechoice); Se SIN printf(* Select the Sourcé Vertex: ee au . scanfl“%d",&s); rms en RT ae ae switch(choice) a { case 1: BFS(s,num); break; case 2: DFS(s,mim); break; . 5| Dwit-4:.Graphs and Sorting f Y to continue else press N> ca scante" wo: while((c etch); } void BFS(int s,int num) //Breath fist search code { int x,i,a5 create(s); Bis}: xedelete(); iffx!=0) printf" %d",x);, while(x!=0) xedeleteQ; iffx!=0) printf(* %d"x); } fori sissnum;i++) it ) BFS(,num); } Void create(int item) _// Create function iff(—=19) Print(“Queue isnot empty”); a t= { gl++1litem; } int delete() Delete function int ks i(n\(=1)) ) if((AtKItI D)&R(Bi}=0) ‘ push_item(i); B[iJ=1; void push_item(int item) /MEunction to perform push operation if(t=19) Stack overflow”); else Stk{+}=item; } {nt pop_item() //Function to perform pop operation. { 8122 4.2 SORTING. 4.2.4 Quick Sort! Q12, Explain in dotail quick sorting method. Provide a complete analysis of quick sort. Answe Model Papers, a8(@) Quick Sort Method Quick sort technique is based on the divide and conquer design technique, In this technique at every step cach element is placed in its proper position, It performs very well on longer lists. I works recursively, by first selecting a random “pivot value” from the list (array). “Then it partitions the list into elements that are fess tha the pivot and. greater than the pivot: The problem of sorting a given list is reduced to the problem of sorting, two sub-lists and process continuous until the list is sorted. This Sorting technique is considered as in place. Algorithm for Quick Sort “This algorithm sortsthesub-array A[fist].ALlast] of a globally defined array A[0....n] in ascending order, Algorithm QuickSort(firsty10s0) Input: Indexes of the elements to be sorted Allast] is a sorted rearrangement oF Output : Affirst, the same elements if{first pivotl high «high = 1 ment) , if\low < high) swap(Allow}, Afhigh}) return high; wr) Algorithm swap(a, b) temp =a aed beter DATA STRUCTURES [JNTU-HYDERABAD) ‘Analysis of Quick Sort “The quicksort is analyzed based on its woist cas, est ease and average case. (Worst-case Analysis: In worstease, the chose, pivot is either the smallest or the largest element inn renay. In this ease, ong partis empty and the other par caniains the remaining clements this generates m~ Jevels in the recursion tree. When quick sort is first called partition takes ime -1, Quick sort is then called twice, once with an empty ray and again with an array of size n~ 1. The next eal {o partition takes time 1 ~2. Quick sort is again called twice, once with an empty array and again with an array of size n—2. Thie next call to partition takes time n-3 nd the process continues, The total time forall the calls to partition n(n) (= I) (= 2) 493) Hae EL if asl) Gil) Best-case Analysis: Quick sort works best if each array is divided into two equal sub-arays of siz= n/2. This generates log n levels in the recursion tree. THe recursion relation is given by, Tn) = 2M!2) + OM) + On applying the master theorem, Tn) = O(n logn) (i) _Avernge-case Analysis: In average-case the array is partitioned by choosing any random number, In this ease at each level some of the partitions well balanced while some are fairly unbalanced. Let us assume that the partition of array to be 9:1 then the recurrence so obtained is, + Ta) = TOn/1O) + Te/0) + Cn ‘The recurrence tree is shown in figure. ciGiven list 4 12, 35; 23, 45, 34, 20, 48] I the elements of given sequence in the sarrangeall | se ‘ofan array, Therefore array (x) } el 35 [23 145 [34 [20 [48 | ype Sette fist element as pivot element, | «pivot = 24 Consider the element next to pivot as i and last } clement as j. Therefore, i= 12,7 = 48 Low High Pele Bs Bs 4s [34 [20 [a8 Pivot i sd ‘1ep3: Since X(i) £ pivot, increment i until the element point by i is greater than pivot. High Te Bs 3 [4s 34 Po [as Pivot i i {: As X(i) > pivot we stop incrementing i and X(j) > pivot decrement j until element X(/) is less than pivot. Low High 24 | 12 [35 | 23 | 45 [34 }20 | 48 Pivot i j As XW) pivot, increment i until element pointed by A (i) is greater than pivot and decrement element pointed by A(/)is less than Pivot. . [2s [iz Teo [23 Fs [34 [35 148 Pivot i B Low High” 24 [12 [20 [23 [45 [34 [35 148 Pivot i 24 [a2 [20 [23 [45 [34 135 148 Pivot (i) = pivot swap pivot a ul j has crossed {he ~ 23 [12 [20 [os ; 3 [12 [20 Daa a5 Baa as Taw Step8: As pivot is placed at its position partition the above list into two sublists, 23 [12 [20 [24 [a5 [94 [as [aa i Pivot 1 ‘Step9: Consider the first sublist S1, Low _" High 23] 1220 Pivot bj Since A() < pivot increment / and j. : 23] 12]20 Pivot i,j ° 23] 1220] . . Pivot ji ‘ Step10: Since A() < pivot and / erossed i, swap AQ) and pivot, 12 [ 23 | 20 | a Foe S, Pivot S, Stepll oo J 12 | 23 | 20 Pivot. i,j 7 Since j cannot be decremented more-and as X(/) pivot stop increnfenting. Now mpare X(i) with pivot since X(/) > pivot decrement j 45 | 34 135 | 48 Pivot J i Step14: Since A{j) < pivot and since / crossed j swap XG) and pivot. 35 [34] 45] 48 Pivot Ss, s, Step15: Since pivots placedatits position divide the above listinto two sublists and then perform sorting. This. can be represented as follows, “st wid as [34 [45] 48 : Pivot hi oer Ar A(j) < pivot swap A(/) and Pivot. Hence 4 is, 3435] 45 | 48 Since the list is already sorted swapping cannot be performed further. Thus the sorted list is as follows, 12 [20 [23 [24 [34 [3s [45 [a8 Justification for “Quick Sort is not a Stable Sorting Method” ‘8 show that quick sort is not a stable sorting algorithm, let ky ky, Ay :.k, ,bea sequence of numbers, A sorting algorithm is stabie if for any two numbers & and k, such that k, = k, and k, precedes &, before sorting (ic., 15). Then k,also proceeds k, after sorting. Quick sort ‘a stable sorting method because the relative order of two equal numbers after sortirig is not same as they were before sort Q13. Apply quick sort to sort thedist E, x, A, M,P, L, E in alphabotical order. Draw the tree of recursive calls made. Answer = sider the given list in the form of an aia, afffownteon el DATA STRUCTURES [INTU-AYDERABAD) Consider the first element ‘E’ as the pivoupy element ic. a (0] = PvtG.e. pivot), and, st the poinan iand jas follows, Dt 2 38 4 SG e)x}alm|ede =| pei i Now, perform quick sort on the above array tg obtain an alphabetically ordered list. Step1:Since, a [i] > pvt(i.e...X > £), iis not incrementeg, However, as a [j] 2 pvt(ic., £2 £),,J is decre. mented as shown below, o 1 2 3 4 5'6 e|xfalmM|]Pp]ele pti j Step2:Since, a [/] again. 7 o 1 2,3 4 5 6 Pat i ‘Step4: Since, a [/] > pvt (i.c., M> E), is decremented o 1 2 3:4 5 6 Ee] x]almM/rp]ocle mm iG StepS:Since, af/] < pvt(i.e., A pvt (i. X> B), therefore, iand/ are not moved. Butthey are exchanged (or swapped) as follows, or 23 4 5 6 Elalxiu| el] e mT Step6:Since, a [i] pvilies X> E), iand jare incremented and decremen ee ee ee Eyalx|[mMieplde4, Grophs.and Sorting cre in no alphabet that comes before vai said to be in correct position, Since E os ater 1A" ftom the given elements, "Eis boat ‘aid to be in correct position. Hence, leav- r a these two elements, comsider the remaining | hements, a8 right sublist | erect position Right tublist der the right sublist and set the first clement as the pivot element. And also set the / and pointers as follows, o 1 2 3 4 5 6 react Thtelx|udelele Lodo 2 Mi i Step$:Since, afi] < pvt, lis incremented. This is contin- ved until the condition is satisfied. o 1 2 3 4 5 6 Gee tarerx[ um] etofe L---- 411 “Pat i epl0: Since, aff] and aff] are less than pivot element. Therefore, the pivot element and the element at Zand j are exchanged as shown below, Cones poston 0 f. 6 totes — asa ‘atelefulel ol x} Lode 23 OT Len vb As there is no alphabet in the array that comes after X; hence it is said to be in correct position. ‘And consider the remaining element as left sub- list ‘PII: Consider the left sublist and set the first element as, the pivot element. And also set the i and j Pointers as follows, oy Jo 3 4 125. Stepl2: Since, Ai]> pili, M> B),sisnot incremented. But, as aff] > priie., L > B)j is decremented. o 1 203 Step13: _ P> E), jis decremented. ‘This is continued until the condition is satisfied. Low High ska sh a |p | L = Right sublist Since, j= pvt and there is no elements that comes before E. Therefore, E is said to be in correct position. Whereas, the remaining elements forms ~ the right sublist. Step14: Now, consider the right sublist and set the first element as the pivot element. And also set the i and j pointers as follows, 22 acs are =e Step15: Since, a{f] > pvt (P > M), iis not incremented. And aff] < pivot(ic., L < Mf), j is not decre- mented. But afi] > aff} (i:e., P< L), therefore - afi] and aff) are swapped. o 1 2,53 4 55 ei Step16: Since, afi] > pvt and alj] < pvt, therefore i and j are not incremented and decremented respectively. However, pivot element M and the element at j are swapped, because £ =o, To transform a ular subtree into a heap, follow the steps given below, (i) Suppose that List{/e] and List(re] are the leftana right children of List{]. (ii) If List[Ic] and List{re] to) find the largest of them. “The index of largest element is stored in “largest. If List] is less than Listflargest], swap Lisc{q] with List[largest]. Otherwise, exit ie., the subtree is already a heap. (iv) | If values have been swapped in step (iii) then, there are chances that the sub-tree Listflarzex] might have become a non-heap. ‘Therefore, we apply steps (ii) and ({ii) at that sub-tree until al the heaps in the sub-tree are restored or until encounter an empty subtree. ‘After constructing a heap, the next phase is, sorting heap begins. Algorithm (Construct Heap) Step 1: Consider a heap tree which is empty. Begin at the first step, i=1 while (i n)do ‘Step 2: Repeat step 2 for all the elements Step 2.1: Select the i* element from list and add = ‘element to the i* place in array n=An{i] Aml[i]=0 i Step 2.2: Repeatstep 2.2 to step 2.3 until the root is checked If Ant{j] > Arr] j/2] then tmp =Arl fj) Arrl{j] = Ari §/2] (ii) End If Step 2.3: End while i=i+l Step 2.4 : End while Step 3: stop - Function a Adjust_maxheap(element any ], int root, it) int ch, root_key; element tmp; tmp = arrfroot]; which is not leaf and has at least one child. root_key = arrfroot] key: ‘Anyone found guilty is LIABLE ta face LEGAL preczedings| | | 4 ora and Sorting ut Tye 2? 100! wwhite(ch <=") i(ch arr[ch].key) breaks else arsfeh/2] = arr{ch}; ch* =2; . ) arr{ch/2] = tmp; Sort Heap In this phase, the given heap is sorted. Note that the root node of heap (ic, the first element of the ist) will be ‘the largest clement in the tree or li following steps sort the heap, 0 t. The ‘wap the root node (i.e, first element of array list) with the last node of the tree (ic, last element of array list). Now, last node is in its proper place. Leave the last element and consider the remaining, elements as the new list. ‘The new list may not be in heap form. Therefore, construct a heap of new list. Gi (iv) Repeat steps (i), (i) and (ii) until all elements are placed at their proper place, ic, until all elements are sorted. Algorithm "Step 1: Read all the elements and store in the array. ‘onstructHeap (Arr) cepeat this step for n — 1 times ‘Step 3.1 Remove the root element and swap it with the ‘element. tmp = Arti] Arrl(i)=Arel[1] Arcl(1] =tmp Step 3.2: Pre build the heap Cannot build heap with on If (=I) then exit Begin at the root node ye element in the tree flag = TRUE i while (flag =TRUE) do 25 right = 24) +41 heck whether heap and check whell move up or not. If (right 05 ‘Adjust_maxheap(arr, i n); forli=n—1;i>0;i--) { SWAP(arr{1], arefi + 1], tmp); Adjust_maxheap(ary, 1, i); 3 3 Example Consider below list oF elerients, tbl DP Dee (ONO CO MO ONO MOMO! Constructing Heap “The first step in heap sorting is to construct heap. For this, consider three pointers i, c, re"where i points to a particular node and fe, re point to the left and right child respectively. Inititly, = S22QE St s Hist i=s-1=4 Left child, e=(2x)t1= 2x4 t1=9 Right child, re=(2x)+2=(2%4)+2=10 Arri (left) 2 Arr [right] then NE JOURNAL FOR ENGINEERING ‘STUDENTS128 8 [7To Tita Se DATA STRUCTURES [JNTU-HYDERABAD} is in the following list, [2 0 PY Br wr ist tr ar aa To i t le aa As, there is no right child of i, the Ic is compared with i. If fe > i, they are swa le they are swapped. Ife i, no swapping In the next iteration, decrement ‘i’ which makes i= 3. Now, Ic and re can be calculated as, i=3 le=(2*3)41=7 re=(2*3)+2=8 é[sl7]°]'[4[3]2]|s5]0 (9) 0) (2) g 4) (5) 1 (7) (8) 19) ae ‘Now compare the element pointed by /e and re (ie., 2 and 5) and find the largest among them. Since 5 > 2, itis compared with / (ie.,.9). Since 9 is greater than 5, the subtree is already a heap, so swapping is not needed. In the next iteration, decrement / which makes it 2. Determine /e and re. sD 14]3]2[s]9] - 4) (5) 16 (7 8) 19) t t Now, compare the elements pointed by le and re (ie., 4 and 3) and find the largest among them. Since 4 >3,4is compared with i(ic., 7). Since 7> 4, the subtree is already a heap, so swapping is not needed. In the next iteration, decrement i which makes it 1. Determine fe and re. i=l le=(2% 1) +1=3 re=(2*1)+2=4 This implies i= 8, le= 9 and re = 1. Now, compare /e and re (ie., 9 and 1) and find the largest among them. Since 9> 1, itis compared with #* element. Since 8 <9, swap these values i.e, i and le. 6 @ 1 2 By 7T8Ti[4]3]2]5]o 2B 6) 1 & Oy Note that the swap operation has distributed the heap in the subtree. So, to restore the heap in that subtree, initialize the pointer i to Ic. . i=3 sle=(2%3)+1=7 re=(2.*3)+2=8 é]>171*1' 1413 [215Je aoa sw 8 6 Mw uy it ‘Swap the largest element among /c and re with i if i 5 which means that the subtree is already heap, so swapping is not needed. In the next iteration, decrement i which makes it .. Determine /e and re. 2 (2x2)+1=5 re=(22)+2=6 This implies i= 7, lc=4 and re=3. Now, compare te and re (i.e., 4 and 3) and compare the largest among Je and re with i(i.e., 4 with 7). Since i> Ic, the subtree is already a heap, so swapping is not needed. ‘The process is repeated for i= 1 For i= 1, there will be no change in the positions of elements. The value of / is decremented again to i=0. i=0 2*0)+1=041 = (2 0)+2=0+2 Now, compare the elements pointed by fe and re (ie., 9 and 7) and compare the largest among them with the value of i. Since 6 <9, the values are swapped. This results in the following list. ® > k > re Q Q ® O 6 © ® OO RIMINAL, t. Anyone fount