Data Structures and Algorithms in Java[1].pdf - Fulvio Frisone

More documents

Recommendations

Info

$(Microsoft PowerPoint - Topic 4.ppt [Mode de compatibilit\351])$

ecursive invocations, and thus the run-time stack, only grows up to size n, because calling factorial(1) returns 1 immediately without invoking itself recursively. The run-time stack allows for method factorial() to exist simultaneously in several active frames (as many as n at some point). Each frame stores the value of its parameter n as well as the value to be returned. Eventually, when the first recursive call terminates, it returns (n − 1)!, which is then multiplied by n to compute n! for the original call of the factorial method. 14.1.2 Allocating Space in the Memory Heap We have already discussed (in Section 14.1.1) how the Java Virtual Machine allocates a method's local variables in that method's frame on the Java run-time stack. The Java stack is not the only kind of memory available for program data in Java, however. Dynamic Memory Allocation Memory for an object can also be allocated dynamically during a method's execution, by having that method utilize the special new operator built into Java. For example, the following Java statement creates an array of integers whose size is given by the value of variable k: int[] items = new int[k]; The size of the array above is known only at runtime. Moreover, the array may continue to exist even after the method that created it terminates. Thus, the memory for this array cannot be allocated on the Java stack. The Memory Heap Instead of using the Java stack for this object's memory, Java uses memory from another area of storage—the memory heap (which should not be confused with the "heap" data structure presented in Chapter 8). We illustrate this memory area, together with the other memory areas, in a Java Virtual Machine in Figure 14.2. The storage available in the memory heap is divided into blocks, which are contiguous array-like "chunks" of memory that may be of variable or fixed sizes. To simplify the discussion, let us assume that blocks in the memory heap are of a fixed size, say, 1,024 bytes, and that one block is big enough for any object we might want to create. (Efficiently handling the more general case is actually an interesting research problem.) Figure 14.2: A schematic view of the layout of memory addresses in the Java Virtual Machine. 890
Memory Allocation Algorithms The Java Virtual Machine definition requires that the memory heap be able to quickly allocate memory for new objects, but it does not specify the data structure that we should use to do this. One popular method is to keep contiguous "holes" of available free memory in a doubly linked list, called the free list. The links joining these holes are stored inside the holes themselves, since their memory is not being used. As memory is allocated and deallocated, the collection of holes in the free lists changes, with the unused memory being separated into disjoint holes divided by blocks of used memory. This separation of unused memory into separate holes is known as fragmentation. Of course, we would like to minimize fragmentation as much as possible. There are two kinds of fragmentation that can occur. Internal fragmentation occurs when a portion of an allocated memory block is not actually used. For example, a program may request an array of size 1000, but only use the first 100 cells of this array. There isn't much that a run-time environment can do to reduce internal fragmentation. External fragmentation, on the other hand, occurs when the there is a significant amount of unused memory between several contiguous blocks of allocated memory. Since the run-time environment has control over where to allocate memory when it is requested (for example, when the new keyword is used in Java), the run-time environment should allocate memory in a way that tries to reduce external fragmentation as much as reasonably possible. Several heuristics have been suggested for allocating memory from the heap so as to minimize external fragmentation. The best-fit algorithm searches the entire free list to find the hole whose size is closest to the amount of memory being requested. The first-fit algorithm searches from the beginning of the free list for the first hole that is large enough. The next-fit algorithm is similar, in that it also searches the free list for the first hole that is large enough, but it begins its search from where it left off previously, viewing the free list as a circularly linked list (Section 3.4.1). The worst-fit algorithm searches the free list to find the largest hole of available memory, which might be done faster than a search of the entire free list if this list were maintained as a priority queue (Chapter 8). In each algorithm, the requested amount of memory is subtracted from the chosen memory hole and the leftover part of that hole is returned to the free list. Although it might sound good at first, the best-fit algorithm tends to produce the worst external fragmentation, since the leftover parts of the chosen holes tend to be small. The first-fit algorithm is fast, but it tends to produce a lot of external 891
Page 1 and 2:
Data Structures and Algorithms in J
Page 3 and 4:
To Isabel -Roberto Tamassia Preface
Page 5 and 6:
The Java code implementing fundamen
Page 7 and 8:
PL6. Object-Oriented Programming Ch
Page 9 and 10:
14. Memory A. Useful Mathematical F
Page 11 and 12:
Sun, Nikos Triandopoulos, Luca Vism
Page 13 and 14:
Getting Started: Classes, Types, an
Page 15 and 16:
39 1.7 An Example Program..........
Page 17 and 18:
The main "actors" in a Java program
Page 19 and 20:
else interface switch boolean exten
Page 21 and 22:
Class modifiers are optional keywor
Page 23 and 24:
Comments Note the use of comments i
Page 25 and 26:
Calling the new operator on a class
Page 27 and 28:
n.longValue( ) float Float n = new
Page 29 and 30:
The Dot Operator Every object refer
Page 31 and 32:
object (assuming we are allowed acc
Page 33 and 34:
Note the uses of instance variables
Page 35 and 36:
1.2 Methods Methods in Java are con
Page 37 and 38:
Abstract methods may only appear wi
Page 39 and 40:
is intended to be used to initializ
Page 41 and 42:
} int i = 512; double e = 2.71828;
Page 43 and 44:
% the modulo operator This last ope
Page 45 and 46:
zeros >>> shift bits right, filling
Page 47 and 48:
3 add./subt. + − 4 shift > >>> 5
Page 49 and 50:
double d1 = 3.2; double d2 = 3.9999
Page 51 and 52:
String s = " " + 4.5; // correct, b
Page 53 and 54:
System.out.println("This is tough."
Page 55 and 56:
Defining a For Loop Here is the syn
Page 57 and 58:
that input and inform the user that
Page 59 and 60:
if (a[i][j] == 0) { foundFlag = tru
Page 61 and 62:
the array, we will sometimes refer
Page 63 and 64:
The cells of this new array, "a," a
Page 65 and 66:
ways to reference such an object. W
Page 67 and 68:
float height = s.nextFloat(); Syste
Page 69 and 70:
We show the CreditCard class in Cod
Page 71 and 72:
Code Fragment 1.7: Output from the
Page 73 and 74:
a text editor class may wish to def
Page 75 and 76:
The process of writing a Java progr
Page 77 and 78:
low-level implementation details. A
Page 79 and 80:
variable declaration, or method def
Page 81 and 82:
• @exception exception-name descr
Page 83 and 84:
A careful testing plan is an essent
Page 85 and 86:
Modify the CreditCard class from Co
Page 87 and 88:
A common punishment for school chil
Page 89 and 90:
Object-Oriented Design Principles .
Page 91 and 92:
java.datastructures.net 2.1 Goals,
Page 93 and 94:
Abstraction The notion of abstracti
Page 95 and 96:
2.1.3 Design Patterns One of the ad
Page 97 and 98:
Example 2.1: Consider a class S tha
Page 99 and 100:
S's a() method when asked for o.a()
Page 101 and 102:
Sometimes, in a Java class, it is c
Page 103 and 104:
103
Page 105 and 106:
Code Fragment 2.3: Class for arithm
Page 107 and 108:
A Fibonacci Progression Class As a
Page 109 and 110:
Figure 2.5 : Inheritance diagram fo
Page 111 and 112:
2.3 Exceptions Exceptions are unexp
Page 113 and 114:
goShopping(); // I don't have to tr
Page 115 and 116:
try problem… // This code might h
Page 117 and 118:
instance, wish to identify some of
Page 119 and 120:
The class BoxedItem shows another f
Page 121 and 122:
comparability feature to a class (i
Page 123 and 124:
• T and S are class types and S i
Page 125 and 126:
We show in Code Fragment 2.13 a cla
Page 127 and 128:
Starting with 5.0, Java includes a
Page 129 and 130:
public void insert (P person, P oth
Page 131 and 132:
• Class Goat extends Object and a
Page 133 and 134:
((Place) usa).printMe(); } } class
Page 135 and 136:
C-2.5 Write a program that consists
Page 137 and 138:
Chapter 3 Arrays, Linked Lists, and
Page 139 and 140:
3.4 Circularly Linked Lists and Lin
Page 141 and 142:
A Class for High Scores Suppose we
Page 143 and 144:
Note that we include a method, toSt
Page 145 and 146:
Once we have identified the place i
Page 147 and 148:
Object Removal Suppose some hot sho
Page 149 and 150:
These methods for adding and removi
Page 151 and 152:
This description is much closer to
Page 153 and 154:
there is no swap needed, and return
Page 155 and 156:
num = [10,12,19,28,33,38,41,46,48,5
Page 157 and 158:
In Code Fragment 3.9, we give a sim
Page 159 and 160:
Many computer games, be they strate
Page 161 and 162:
We give a complete Java class for m
Page 163 and 164:
Code Fragment 3.11: A simple, compl
Page 165 and 166:
Like an array, a singly linked list
Page 167 and 168:
Code Fragment 3.14: Inserting a new
Page 169 and 170:
The reverse operation of inserting
Page 171 and 172:
Header and Trailer Sentinels To sim
Page 173 and 174:
Likewise, we can easily perform an
Page 175 and 176:
Figure 3.17: Adding a new node afte
Page 177 and 178:
Code Fragment 3.23: Java class DLis
Page 179 and 180:
We make the following observations
Page 181 and 182:
advance(): Advance the cursor to th
Page 183 and 184:
Some Observations about the CircleL
Page 185 and 186:
Some Sample Output 185
Page 187 and 188:
We show in Code Fragment 3.27 thein
Page 189 and 190:
The factorial function can be defin
Page 191 and 192:
Each multiple of 1 inch also has a
Page 193 and 194:
The trace for drawTicks is more com
Page 195 and 196:
The simplest form of recursion is l
Page 197 and 198:
nonrecursive part of each call. Mor
Page 199 and 200:
3.5.2 Binary Recursion When an algo
Page 201 and 202:
Unfortunately, in spite of the Fibo
Page 203 and 204:
To solve such a puzzle, we need to
Page 205 and 206:
3.6 Exercises For source code and h
Page 207 and 208:
C-3.2 Let A be an array of size n
Page 209 and 210:
Write a recursive Java program that
Page 211 and 212:
Chapter Notes The fundamental data
Page 213 and 214:
Primitive Operations . . . . . . .
Page 215 and 216:
4.1.2 The Logarithm function One of
Page 217 and 218:
4.1.5 The Quadratic function Anothe
Page 219 and 220:
which assigns to an input value n t
Page 221 and 222:
For example, we have the following:
Page 223 and 224:
primarily as slopes. Even so, the e
Page 225 and 226:
and y-coordinate equal to the runni
Page 227 and 228:
Counting Primitive Operations Speci
Page 229 and 230:
Simplifying the Analysis Further In
Page 231 and 232:
Proposition 4.7: The Algorithmarray
Page 233 and 234:
is O(4n 2 + 6n log n), although thi
Page 235 and 236:
256 16 4 16 64 256 4,096 65,536 32
Page 237 and 238:
Running Maximum Problem Size (n) Ti
Page 239 and 240:
should the constant factors they "h
Page 241 and 242:
• There are two nested for loops,
Page 243 and 244:
This definition leads immediately t
Page 245 and 246:
Besides showing a use of the contra
Page 247 and 248:
We may sometimes feel overwhelmed b
Page 249 and 250:
The number of operations executed b
Page 251 and 252:
R-4.22 Show that if d(n) is O(f(n))
Page 253 and 254:
R-4.29 Show that 2 n+1 is O(2 n ).
Page 255 and 256:
Describe a method for finding both
Page 257 and 258:
Describe, in pseudo-code a method f
Page 259 and 260:
Stacks.............................
Page 261 and 262:
down on the stack to become the new
Page 263 and 264:
7 (5) top() 5 (5) pop() 5 () pop()
Page 265 and 266:
In addition, we must define excepti
Page 267 and 268:
5.1.2 A Simple Array-Based Stack Im
Page 269 and 270:
we indicate that the stack no longe
Page 271 and 272:
271
Page 273 and 274:
273
Page 275 and 276:
The array implementation of a stack
Page 277 and 278:
of the list and return its element.
Page 279 and 280:
5.1.5 Matching Parentheses and HTML
Page 281 and 282:
Matching Tags in an HTML Document A
Page 283 and 284:
Code Fragment 5.12: Java program fo
Page 285 and 286:
Another fundamental data structure
Page 287 and 288:
true ( ) enqueue(9) - (9) enqueue(7
Page 289 and 290:
5.2.2 A Simple Array-Based Queue Im
Page 291 and 292:
Using the Modulo Operator to Implem
Page 293 and 294:
which gives the correct result both
Page 295 and 296:
Each of the methods of the singly l
Page 297 and 298:
potato is equivalent to dequeuing a
Page 299 and 300:
- (3) addFirst(5) - (5,3) removeFir
Page 301 and 302:
Thus, a doubly linked list can be u
Page 303 and 304:
Code Fragment 5.18: Class NodeDeque
Page 305 and 306:
Page 307 and 308:
Suppose Alice has picked three dist
Page 309 and 310:
Implement a program that can input
Page 311 and 312:
223 6.1.3 A Simple Array-Based Impl
Page 313 and 314:
Sequences....................... 25
Page 315 and 316:
- (7) add(0,4) - (4,7) get(1) 7 (4,
Page 317 and 318:
get(size() −1) addFirst(e) add(0,
Page 319 and 320:
have to be shifted forward. A simil
Page 321 and 322:
Moreover, the java.util.ArrayList c
Page 323 and 324:
An Amortized Analysis of Extendable
Page 325 and 326:
Proposition 6.2: Let S be an array
Page 327 and 328:
linked list without revealing this
Page 329 and 330:
The node list ADT allows us to view
Page 331 and 332:
(8,3,5) set(p 3 ,7) 3 (8,7,5) addAf
Page 333 and 334:
Deque Method Realization with Node-
Page 335 and 336:
new node v to hold the element e, l
Page 337 and 338:
In conclusion, using a doubly linke
Page 339 and 340:
Code Fragment 6.11: Portions of the
Page 341 and 342:
An iterator is a software design pa
Page 343 and 344:
Since looping through the elements
Page 345 and 346:
Code Fragment 6.15: The iterator()
Page 347 and 348:
the last element. That is, it uses
Page 349 and 350:
isEmpty() isEmpty() O(1) time get(i
Page 351 and 352:
deletion is at cursor; A is O(n),L
Page 353 and 354:
where n is the number of elements i
Page 355 and 356:
Efficiency Trade-Offs with an Array
Page 357 and 358:
357
Page 359 and 360:
The previous implementation of a fa
Page 361 and 362:
decreasing order and then pops up a
Page 363 and 364:
accesses several Web pages and then
Page 365 and 366:
Page 367 and 368:
R-6.16 Describe how to create an it
Page 369 and 370:
Think about how to extend the circu
Page 371 and 372:
C-6.24 Isabel has an interesting wa
Page 373 and 374:
Chapter 7 Trees Contents 7.1 Genera
Page 375 and 376:
7.3.3 Properties of Binary Trees ..
Page 377 and 378:
Electronics R'Us. The children of t
Page 379 and 380:
Edges and Paths in Trees An edge of
Page 381 and 382:
There are also a number of generic
Page 383 and 384:
Figure 7.5: The linked structure fo
Page 385 and 386:
Code Fragment 7.3: Method depth wri
Page 387 and 388:
Code Fragment 7.6: Algorithm height
Page 389 and 390:
Example 7.6: The preorder traversal
Page 391 and 392:
Code Fragment 7.10: Algorithm paren
Page 393 and 394:
The postorder traversal method is u
Page 395 and 396:
Although the preorder and postorder
Page 397 and 398:
A Recursive Binary Tree Definition
Page 399 and 400:
same depth d as the level dof T. In
Page 401 and 402:
Figure 7.13: Operation that removes
Page 403 and 404:
We use a Java interface BTPosition
Page 405 and 406:
addition to the Binary Tree interfa
Page 407 and 408:
Page 409 and 410:
Page 411 and 412:
Binary Tree interface. (Continues i
Page 413 and 414:
• Methods size() and isEmpty() us
Page 415 and 416:
The level numbering function p sugg
Page 417 and 418:
Time size, isEmpty O(1) iterator, p
Page 419 and 420:
As is the case for general trees, t
Page 421 and 422:
The inorder traversal of a binary t
Page 423 and 424:
The inorder traversal can also be a
Page 425 and 426:
start an Euler tour by initializing
Page 427 and 428:
• Calls auxiliary method visitRig
Page 429 and 430:
ExpressionTerm at each node. Class
Page 431 and 432:
Code Fragment 7.31: Class EvaluateE
Page 433 and 434:
Let T be an n-node binary tree that
Page 435 and 436:
c. Let T be a proper binary tree wi
Page 437 and 438:
Creativity C-7.1 For each node v in
Page 439 and 440:
C-7.10 Two ordered trees T ′ and
Page 441 and 442:
postorderDraw is similar to preorde
Page 443 and 444:
Describe a nonrecursive method for
Page 445 and 446:
P-7.8 Write a program that can play
Page 447 and 448:
Chapter 8 Priority Queues Contents
Page 449 and 450:
The Heap Data Structure............
Page 451 and 452:
As in the examples above from an ai
Page 453 and 454:
package includes a similar Entry in
Page 455 and 456:
Code Fragment 8.3: Class representi
Page 457 and 458:
insert(5,A) e 1 [=(5,A)] {(5,A)} in
Page 459 and 460:
size() size() isEmpty() isEmpty() i
Page 461 and 462:
at the time the method is executed.
Page 463 and 464:
Code Fragment 8.7: Java class Defau
Page 465 and 466:
465
Page 467 and 468:
Insertion-Sort If we implement the
Page 469 and 470:
An efficient realization of a prior
Page 471 and 472:
Proposition 8.5: A heap T storing n
Page 473 and 474:
The Array List Representation of a
Page 475 and 476:
java.util.ArrayList. (Continues in
Page 477 and 478:
Code Fragment 8.12: Class ArrayList
Page 479 and 480:
8.3.3 Implementing a Priority Queue
Page 481 and 482:
The upward movement of the newly in
Page 483 and 484:
Analysis Table 8.3 shows the runnin
Page 485 and 486:
We conclude that the heap data stru
Page 487 and 488:
Code Fragment 8.15: Remaining auxil
Page 489 and 490:
As we have previously observed, rea
Page 491 and 492:
8.3.6 Bottom-Up Heap Construction
Page 493 and 494:
Recursive Bottom-Up Heap Constructi
Page 495 and 496:
Figure 8.11: Visual justification o
Page 497 and 498:
insert(3,B) e 2 {(3,B), (5,A)} inse
Page 499 and 500:
The performance of an adaptable pri
Page 501 and 502:
Code Fragment 8.18: An adaptable pr
Page 503 and 504:
503
Page 505 and 506:
Illustrate the execution of the sel
Page 507 and 508:
A group of children want to play a
Page 509 and 510:
C-8.13 We can represent a path from
Page 511 and 512:
P-8.1 Give a Java implementation of
Page 513 and 514:
Chapter 9 Maps and Dictionaries Con
Page 515 and 516:
9.3.3 Ordered Search Tables and Bin
Page 517 and 518:
In either case, we use a key as a u
Page 519 and 520:
null {(5,A), (7,B)} put(2,C) null {
Page 521 and 522:
get(k) put(k,v) put(k,v) remove(k)
Page 523 and 524:
time of map operations in an n-entr
Page 525 and 526:
Hash Codes in Java The generic Obje
Page 527 and 528:
of the word lists provided in two v
Page 529 and 530:
6 6 2 7 14 2 8 105 2 9 18 2 10 277
Page 531 and 532:
however, for if there is a repeated
Page 533 and 534:
only the one entry. We leave the de
Page 535 and 536:
key k will skip over cells containi
Page 537 and 538:
537
Page 539 and 540:
539
Page 541 and 542:
9.2.7 Load Factors and Rehashing In
Page 543 and 544:
example, when pundits study politic
Page 545 and 546:
for the keys, and we provide additi
Page 547 and 548:
findAll(2) {(2,C), (2,E)} {(5,A), (
Page 549 and 550:
Analysis of a List-Based Dictionary
Page 551 and 552:
for entries with duplicate keys. As
Page 553 and 554:
Figure 9.7: Realization of a dictio
Page 555 and 556:
Considering the running time of bin
Page 557 and 558:
Method List Hash Table Search Table
Page 559 and 560:
An example of a skip list is shown
Page 561 and 562:
We give a pseudo-code description o
Page 563 and 564:
the new entry are drawn with thick
Page 565 and 566:
can go into what is almost an infin
Page 567 and 568:
of entries in the dictionary at the
Page 569 and 570:
O(n) O(logn) (expected) 9.5.2 The O
Page 571 and 572:
lastKey() last().getValue() get(las
Page 573 and 574:
Formally, we say a price-performanc
Page 575 and 576:
What is the worst-case running time
Page 577 and 578:
Describe how to use a map to implem
Page 579 and 580:
in memory, describe a method runnin
Page 581 and 582:
please see the various research pap
Page 583 and 584:
Amortized Analysis of Splaying ★.
Page 585 and 586:
10.1.1 Searching To perform operati
Page 587 and 588:
We can also show that a variation o
Page 589 and 590:
The implementation of the remove(k)
Page 591 and 592:
O(h) time, where h is the height of
Page 593 and 594:
Class BinarySearchTree uses locatio
Page 595 and 596:
Code Fragment 10.4: Class BinarySea
Page 597 and 598:
Code Fragment 10.5: Class BinarySea
Page 599 and 600:
10.2 AVL Trees In the previous sect
Page 601 and 602:
(10.3) > 2 i · n(h − 2i). That i
Page 603 and 604:
We restore the balance of the nodes
Page 605 and 606:
Figure 10.9: Schematic illustration
Page 607 and 608:
find another, we perform a restruct
Page 609 and 610:
In Code Fragments 10.7 and 10.8, we
Page 611 and 612:
Code Fragment 10.8: Auxiliary metho
Page 613 and 614:
10.3 Splay Trees Another way we can
Page 615 and 616:
We perform a zig-zig or a zig-zag w
Page 617 and 618:
Figure 10.16: Example of splaying a
Page 619 and 620:
10.3.2 When to Splay The rules that
Page 621 and 622:
10.3.3 Amortized Analysis of Splayi
Page 623 and 624:
When we perform a splaying, we pay
Page 625 and 626:
≤ r ′(x)−r(x) ≤ 3(r ′(x)
Page 627 and 628:
Proposition 10.6: Consider a sequen
Page 629 and 630:
629
Page 631 and 632:
The dictionary we store at each nod
Page 633 and 634:
h≤log(n+1) (10.10) are true. To j
Page 635 and 636:
insertion of 5, which causes an ove
Page 637 and 638:
A split operation affects a constan
Page 639 and 640:
operation; (f) after the fusion ope
Page 641 and 642:
collection returned by findAll. The
Page 643 and 644:
Figure 10.27: Correspondence betwee
Page 645 and 646:
• Take node z, its parent v, and
Page 647 and 648:
Figures 10.30 and 10.31 show a sequ
Page 649 and 650:
The cases for insertion imply an in
Page 651 and 652:
and (b) and for node b in part (c)
Page 653 and 654:
Case 2: The Sibling y of r is Black
Page 655 and 656:
Case 3: The Sibling y of r is Red.
Page 657 and 658:
constant amount of work (in a restr
Page 659 and 660:
Performance of Red-Black Trees Tabl
Page 661 and 662:
Thus, a red-black tree achieves log
Page 663 and 664:
Methods insert (Code Fragment 10.10
Page 665 and 666:
665
Page 667 and 668:
Are the rotations in Figures 10.8 a
Page 669 and 670:
T is a (2,4) tree. c. T is a red-bl
Page 671 and 672:
C-10.8 Let D be an ordered dictiona
Page 673 and 674:
Projects P-10.1 N-body simulations
Page 675 and 676:
early approaches to balancing trees
Page 677 and 678:
11.5 Comparison of Sorting Algorith
Page 679 and 680:
elements of S; that is, S 1 contain
Page 681 and 682:
for a child invocation to return. (
Page 683 and 684:
Proposition 11.1: The merge-sort tr
Page 685 and 686:
idea is to iteratively remove the s
Page 687 and 688:
merge-sort, are implemented by eith
Page 689 and 690:
Code Fragment 11.3: Methods mergeSo
Page 691 and 692:
A Nonrecursive Array-Based Implemen
Page 693 and 694:
the case when n is a power of 2. (W
Page 695 and 696:
each node of the quick-sort tree. T
Page 697 and 698:
eturn). Note the divide steps perfo
Page 699 and 700:
Performing Quick-Sort on Arrays and
Page 701 and 702:
E has at least one element (the piv
Page 703 and 704:
the size of the input for this call
Page 705 and 706:
In-place quick-sort modifies the in
Page 707 and 708:
Unfortunately, the implementation a
Page 709 and 710:
algorithm to output the elements of
Page 711 and 712:
itself is a sequence (of entries wi
Page 713 and 714:
first component, we guarantee that
Page 715 and 716:
Bucket-Sort and Radix-Sort Finally,
Page 717 and 718:
generic merging takes time proporti
Page 719 and 720:
• In class Union Merge, merge cop
Page 721 and 722:
Performance of the Sequence Impleme
Page 723 and 724:
With this partition data structure,
Page 725 and 726:
The Log-Star and Inverse Ackermann
Page 727 and 728:
This may come as a small surprise,
Page 729 and 730:
is the same as the expected number
Page 731 and 732:
Show that the probability that any
Page 733 and 734:
Describe what additional informatio
Page 735 and 736:
Argue why the previous claim proves
Page 737 and 738:
threads of b. Describe and analyze
Page 739 and 740:
Design and implement a stable versi
Page 741 and 742:
Contents 12.1 String Operations ...
Page 743 and 744:
567 12.5 Text Similarity Testing ..
Page 745 and 746:
finite, we usually assume that the
Page 747 and 748:
Return and update S to be the strin
Page 749 and 750:
Performance The brute-force pattern
Page 751 and 752:
We will formalize these heuristics
Page 753 and 754:
In Figure 12.3, we illustrate the e
Page 755 and 756:
A Java implementation of the BM pat
Page 757 and 758:
using an alternative shift heuristi
Page 759 and 760:
In Figure 12.5, we illustrate the e
Page 761 and 762:
A Java implementation of the KMP pa
Page 763 and 764:
12.3 Tries The pattern matching alg
Page 765 and 766:
The worst case for the number of no
Page 767 and 768:
A compressed trie is similar to a s
Page 769 and 770:
trie. In the next section, we prese
Page 771 and 772:
if such a path can be traced. The d
Page 773 and 774:
Let m be the size of pattern P and
Page 775 and 776:
transmit our text. Likewise, text c
Page 777 and 778:
12.4.2 The Greedy Method Huffman's
Page 779 and 780:
There are few algorithmic technique
Page 781 and 782:
Turning this definition of L[i, j]
Page 783 and 784:
Reinforcement R-12.1 List the prefi
Page 785 and 786:
C-12.5 Say that a pattern P of leng
Page 787 and 788:
suffix of length j in y. Design an
Page 789 and 790:
pattern matching. The reader intere
Page 791 and 792:
Depth-First Search Traversing a Dig
Page 793 and 794:
Example 13.2: We can associate with
Page 795 and 796:
In the propositions that follow, we
Page 797 and 798:
are communication connections betwe
Page 799 and 800:
vertices and m edges, an edge list
Page 801 and 802:
The main feature of the edge list s
Page 803 and 804:
• The edge object for an edge e w
Page 805 and 806:
insertVertex, insertEdge, removeEdg
Page 807 and 808:
not. The optimal performance of met
Page 809 and 810:
came from to get to v, and repeat t
Page 811 and 812:
We can visualize a DFS traversal by
Page 813 and 814:
• Computing a cycle in G, or repo
Page 815 and 816:
depend on the implementation of the
Page 817 and 818:
Code Fragment 13.5: The main templa
Page 819 and 820:
Using the Template Method Pattern f
Page 821 and 822:
Class FindCycleDFS (Code Fragment 1
Page 823 and 824:
levels. BFS can also be thought of
Page 825 and 826:
825
Page 827 and 828:
• Given a start vertex s of G, co
Page 829 and 830:
Interesting problems that deal with
Page 831 and 832:
A DFS on a digraph partitions the e
Page 833 and 834:
Directed Breadth-First Search As wi
Page 835 and 836:
edges (vi,vk) and (vk,vj) with dash
Page 837 and 838:
ordering such that any directed pat
Page 839 and 840: Justification: The initial computat
Page 841 and 842: computers. In this case, it is prob
Page 843 and 844: their distances from v. Thus, in ea
Page 845 and 846: edge for pulling in u with a solid
Page 847 and 848: Proposition 13.23: In Dijkstra's al
Page 849 and 850: Referring back to Code Fragment 13.
Page 851 and 852: The main computation of Dijkstra's
Page 853 and 854: 13.7 Minimum Spanning Trees Suppose
Page 855 and 856: 13.7.1 Kruskal's Algorithm The reas
Page 857 and 858: Figure 13.19: An example of an exec
Page 859 and 860: execution of Kruskal's algorithm. (
Page 861 and 862: Analyzing the Prim-Jarn ık Algorit
Page 863 and 864: 13.8 Exercises For source code and
Page 865 and 866: LA22: (none) • LA31: LA15 • LA3
Page 867 and 868: The graph has 10,000 vertices and 2
Page 869 and 870: 265 175 215 180 185 155 3- - - - 26
Page 871 and 872: _ _ _ _ Find which bridges to build
Page 873 and 874: determine if it is constructed corr
Page 875 and 876: An Euler tour of a directed graph t
Page 877 and 878: C-13.21 Computer networks should av
Page 879 and 880: c. Now suppose that is weighted and
Page 881 and 882: P-13.9 Design an experimental compa
Page 883 and 884: Chapter 14 Memory Contents 14.1 Mem
Page 885 and 886: Multi-way Merging..................
Page 887 and 888: Keeping Track of the Program Counte
Page 889: In Section 7.3.6 we described how t
Page 893 and 894: identify and mark each live object.
Page 895 and 896: Caches and Disks Specifically, the
Page 897 and 898: are addressed. Virtual memory does
Page 899 and 900: The Random strategy is one of the e
Page 901 and 902: each would require to perform the s
Page 903 and 904: describe this structure, however, l
Page 905 and 906: are too large to fit entirely into
Page 907 and 908: Reinforcement R-14.1 Describe, in d
Page 909 and 910: Describe how to use a B-tree to imp
Page 911 and 912: Appendix A Useful Mathematical Fact
Page 913 and 914: The factorial function is defined a
Page 915 and 916: for any real number 0 < a ≠ 1. Pr
Page 917 and 918: E(X + Y) = E(X) + E(Y) and E(cX) =
Page 919 and 920: Bibliography [1] G. M. Adelson-Vels
Page 921 and 922: [28] S. A. Demurjian, Sr., Software
Page 923 and 924: [69] B. Liskov and J. Guttag, Abstr
show all

Data Structures and Algorithms in Java[1].pdf - Fulvio Frisone

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?