Source

fcla / src / section-O.xml

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
<?xml version="1.0" encoding="UTF-8" ?>
<section acro="O">
<title>Orthogonality</title>

<!-- %%%%%%%%%% -->
<!-- % -->
<!-- %  Section O -->
<!-- %  Orthogonality -->
<!-- % -->
<!-- %%%%%%%%%% -->
<introduction>
<p>In this section we define a couple more operations with vectors, and prove a few theorems.  At first blush these definitions and results will not appear central to what follows, but we will make use of them at key points in the remainder of the course (such as <acroref type="section" acro="MINM" />, <acroref type="section" acro="OD" />).  Because we have chosen to use $\complexes$ as our set of scalars, this subsection is a bit more, uh, <ellipsis /> complex than it would be for the real numbers.  We'll explain as we go along how things get easier for the real numbers ${\mathbb R}$.  If you haven't already, now would be a good time to review some of the basic properties of arithmetic with complex numbers described in <acroref type="section" acro="CNO" />.  With that done, we can extend the basics of complex number arithmetic to our study of vectors in $\complex{m}$.</p>

</introduction>

<subsection acro="CAV">
<title>Complex Arithmetic and Vectors</title>

<p>We know how the addition and multiplication of complex numbers is employed in defining the operations for vectors in $\complex{m}$ (<acroref type="definition" acro="CVA" /> and <acroref type="definition" acro="CVSM" />).  We can also extend the idea of the conjugate to vectors.</p>

<definition acro="CCCV" index="conjugate!column vector">
<title>Complex Conjugate of a Column Vector</title>
<p>Suppose that $\vect{u}$ is a vector from $\complex{m}$.  Then the conjugate of the vector, $\conjugate{\vect{u}}$, is defined by
<alignmath>
\vectorentry{\conjugate{\vect{u}}}{i}
<![CDATA[&=\conjugate{\vectorentry{\vect{u}}{i}}]]>
<![CDATA[&&\text{$1\leq i\leq m$}]]>
</alignmath>
</p>

<notation acro="CCCV" index="conjugate of a vector">
<title>Complex Conjugate of a Column Vector</title>
<usage>$\conjugate{\vect{u}}$</usage>
</notation>
</definition>

<p>With this definition we can show that the conjugate of a column vector behaves as we would expect with regard to vector addition and scalar multiplication.</p>

<theorem acro="CRVA" index="conjugate!vector addition">
<title>Conjugation Respects Vector Addition</title>
<statement>
<p>Suppose $\vect{x}$ and $\vect{y}$ are two vectors from $\complex{m}$.  Then
<equation>
\conjugate{\vect{x}+\vect{y}}=\conjugate{\vect{x}}+\conjugate{\vect{y}}
</equation>
</p>

</statement>

<proof>
<p>For each $1\leq i\leq m$,
<alignmath>
\vectorentry{\conjugate{\vect{x}+\vect{y}}}{i}
<![CDATA[&=\conjugate{\vectorentry{\vect{x}+\vect{y}}{i}}&&]]>\text{<acroref type="definition" acro="CCCV" />}\\
<![CDATA[&=\conjugate{\vectorentry{\vect{x}}{i}+\vectorentry{\vect{y}}{i}}&&]]>\text{<acroref type="definition" acro="CVA" />}\\
<![CDATA[&=\conjugate{\vectorentry{\vect{x}}{i}}+\conjugate{\vectorentry{\vect{y}}{i}}&&]]>\text{<acroref type="theorem" acro="CCRA" />}\\
<![CDATA[&=\vectorentry{\conjugate{\vect{x}}}{i}+\vectorentry{\conjugate{\vect{y}}}{i}&&]]>\text{<acroref type="definition" acro="CCCV" />}\\
<![CDATA[&=\vectorentry{\conjugate{\vect{x}}+\conjugate{\vect{y}}}{i}&&]]>\text{<acroref type="definition" acro="CVA" />}
</alignmath>
</p>

<p>Then by <acroref type="definition" acro="CVE" /> we have $\conjugate{\vect{x}+\vect{y}}=\conjugate{\vect{x}}+\conjugate{\vect{y}}$.</p>

</proof>
</theorem>

<theorem acro="CRSM" index="conjugate!scalar multiplication">
<title>Conjugation Respects Vector Scalar Multiplication</title>
<statement>
<p>Suppose $\vect{x}$ is a vector from $\complex{m}$, and $\alpha\in\complexes$ is a scalar.  Then
<equation>
\conjugate{\alpha\vect{x}}=\conjugate{\alpha}\,\conjugate{\vect{x}}
</equation>
</p>

</statement>

<proof>
<p>For $1\leq i\leq m$,
<alignmath>
\vectorentry{\conjugate{\alpha\vect{x}}}{i}
<![CDATA[&=\conjugate{\vectorentry{\alpha\vect{x}}{i}}&&]]>\text{<acroref type="definition" acro="CCCV" />}\\
<![CDATA[&=\conjugate{\alpha\vectorentry{\vect{x}}{i}}&&]]>\text{<acroref type="definition" acro="CVSM" />}\\
<![CDATA[&=\conjugate{\alpha}\,\conjugate{\vectorentry{\vect{x}}{i}}&&]]>\text{<acroref type="theorem" acro="CCRM" />}\\
<![CDATA[&=\conjugate{\alpha}\,\vectorentry{\conjugate{\vect{x}}}{i}&&]]>\text{<acroref type="definition" acro="CCCV" />}\\
<![CDATA[&=\vectorentry{\conjugate{\alpha}\,\conjugate{\vect{x}}}{i}&&]]>\text{<acroref type="definition" acro="CVSM" />}\\
</alignmath>
</p>

<p>Then by <acroref type="definition" acro="CVE" /> we have $\conjugate{\alpha\vect{x}}=\conjugate{\alpha}\,\conjugate{\vect{x}}$.</p>

</proof>
</theorem>

<p>These two theorems together tell us how we can <q>push</q> complex conjugation through linear combinations.
</p>

</subsection>

<subsection acro="IP">
<title>Inner products</title>

<definition acro="IP" index="vector!inner product">
<title>Inner Product</title>
<p>Given the vectors $\vect{u},\,\vect{v}\in\complex{m}$ the <define>inner product</define> of $\vect{u}$ and $\vect{v}$ is the scalar quantity in $\complex{\null}$,
<equation>
\innerproduct{\vect{u}}{\vect{v}}=
\conjugate{\vectorentry{\vect{u}}{1}}\vectorentry{\vect{v}}{1}+
\conjugate{\vectorentry{\vect{u}}{2}}\vectorentry{\vect{v}}{2}+
\conjugate{\vectorentry{\vect{u}}{3}}\vectorentry{\vect{v}}{3}+
\cdots+
\conjugate{\vectorentry{\vect{u}}{m}}\vectorentry{\vect{v}}{m}
=
\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{v}}{i}
</equation>
</p>

<notation acro="IP" index="inner product">
<title>Inner Product</title>
<usage>$\innerproduct{\vect{u}}{\vect{v}}$</usage>
</notation>
</definition>

<p>This operation is a bit different in that we begin with two vectors but produce a scalar.  Computing one is straightforward.
</p>

<example acro="CSIP" index="inner product">
<title>Computing some inner products</title>

<p>The inner product of
<alignmath>
<![CDATA[\vect{u}=\colvector{2+3i\\5+2i\\-3+i}&&\text{and}&&]]>
\vect{v}=\colvector{1+2i\\-4+5i\\0+5i}
</alignmath>
is
<alignmath>
\innerproduct{\vect{u}}{\vect{v}}
<![CDATA[&=(\conjugate{2+3i})(1+2i)+(\conjugate{5+2i})(-4+5i)+(\conjugate{-3+i})(0+5i)\\]]>
<![CDATA[&=(2-3i)(1+2i)+(5-2i)(-4+5i)+(-3-i)(0+5i)\\]]>
<![CDATA[&=(8+i)+(-10+33i)+(5-15i)\\]]>
<![CDATA[&=3+19i]]>
</alignmath>
</p>

<p>The inner product of
<alignmath>
<![CDATA[\vect{w}=\colvector{2\\4\\-3\\2\\8}&&\text{and}&&]]>
\vect{x}=\colvector{3\\1\\0\\-1\\-2}
</alignmath>
is
<alignmath>
<![CDATA[\innerproduct{\vect{w}}{\vect{x}}&=]]>
(\conjugate{2})3+(\conjugate{4})1+(\conjugate{-3})0+(\conjugate{2})(-1)+(\conjugate{8})(-2)\\
<![CDATA[&=2(3)+4(1)+(-3)0+2(-1)+8(-2)=-8.]]>
</alignmath>
</p>

</example>

<p>In the case where the entries of our vectors are all real numbers (as in the second part of <acroref type="example" acro="CSIP" />), the computation of the inner product may look familiar and be known to you as a <define>dot product</define> or <define>scalar product</define>.  So you can view the inner product as a generalization of the scalar product to vectors from $\complex{m}$ (rather than ${\mathbb R}^m$).</p>

<p>Note that we have chosen to conjugate the entries of the <em>first</em> vector listed in the inner product, while it is almost equally feasible to conjugate entries from the <em>second</em> vector instead.  In particular, prior to Version 2.90, we did use the latter definition, and this has now changed to the former, with resulting adjustments propogated up through <acroref type="section" acro="CB" /> (only).  However, conjugating the first vector leads to much nicer formulas for certain matrix decompositions and also shortens some proofs.</p>

<p>There are several quick theorems we can now prove, and they will each be useful later.</p>

<theorem acro="IPVA" index="inner product!vector addition">
<title>Inner Product and Vector Addition</title>
<statement>
<p>Suppose $\vect{u},\,\vect{v},\,\vect{w}\in\complex{m}$.  Then
<ol>
<li>$\innerproduct{\vect{u}+\vect{v}}{\vect{w}}=\innerproduct{\vect{u}}{\vect{w}}+\innerproduct{\vect{v}}{\vect{w}}$</li>
<li>$\innerproduct{\vect{u}}{\vect{v}+\vect{w}}=\innerproduct{\vect{u}}{\vect{v}}+\innerproduct{\vect{u}}{\vect{w}}$</li>
</ol>
</p>

</statement>

<proof>
<p>The proofs of the two parts are very similar, with the second one requiring just a bit more effort due to the conjugation that occurs.  We will prove part 1 and you can prove part 2 (<acroref type="exercise" acro="O.T10" />).
<alignmath>
\innerproduct{\vect{u}+\vect{v}}{\vect{w}}
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}+\vect{v}}{i}}\vectorentry{\vect{w}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}\\
<![CDATA[&=\sum_{i=1}^{m}\left(\conjugate{\vectorentry{\vect{u}}{i}+]]>
\vectorentry{\vect{v}}{i}}\right)\vectorentry{\vect{w}}{i}
<![CDATA[&&]]>\text{<acroref type="definition" acro="CVA" />}\\
<![CDATA[&=\sum_{i=1}^{m}\left(\conjugate{\vectorentry{\vect{u}}{i}}+]]>
\conjugate{\vectorentry{\vect{v}}{i}}\right)\vectorentry{\vect{w}}{i}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="CCRA" />}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{w}}{i}]]>
              + \conjugate{\vectorentry{\vect{v}}{i}}\vectorentry{\vect{w}}{i}
<![CDATA[&&]]>\text{<acroref type="property" acro="DCN" />}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{w}}{i}]]>
 +\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{v}}{i}}\vectorentry{\vect{w}}{i}
<![CDATA[&&]]>\text{<acroref type="property" acro="CACN" />}\\
<![CDATA[&=\innerproduct{\vect{u}}{\vect{w}}+\innerproduct{\vect{v}}{\vect{w}}&&]]>\text{<acroref type="definition" acro="IP" />}
</alignmath>
</p>

</proof>
</theorem>

<theorem acro="IPSM" index="inner product!scalar multiplication">
<title>Inner Product and Scalar Multiplication</title>
<statement>
<p>Suppose $\vect{u},\,\vect{v}\in\complex{m}$ and $\alpha\in\complex{\null}$.  Then
<ol>
<li>$\innerproduct{\alpha\vect{u}}{\vect{v}}=\conjugate{\alpha}\innerproduct{\vect{u}}{\vect{v}}$</li>
<li>$\innerproduct{\vect{u}}{\alpha\vect{v}}=\alpha\innerproduct{\vect{u}}{\vect{v}}$</li>
</ol>
</p>

</statement>

<proof>
<p>The proofs of the two parts are very similar, with the second one requiring just a bit more effort due to the conjugation that occurs.  We will prove part 1 and you can prove part 2  (<acroref type="exercise" acro="O.T11" />).
<alignmath>
\innerproduct{\alpha\vect{u}}{\vect{v}}
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\alpha\vect{u}}{i}}\vectorentry{\vect{v}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\alpha\vectorentry{\vect{u}}{i}}\vectorentry{\vect{v}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="CVSM" />}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\alpha}\,\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{v}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="theorem" acro="CCRM" />}\\
<![CDATA[&=\conjugate{\alpha}\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{v}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="property" acro="DCN" />}\\
<![CDATA[&=\conjugate{\alpha}\innerproduct{\vect{u}}{\vect{v}}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}
</alignmath>
</p>

</proof>
</theorem>

<theorem acro="IPAC" index="inner product!anti-commutative">
<title>Inner Product is Anti-Commutative</title>
<statement>
<p>Suppose that $\vect{u}$ and $\vect{v}$ are vectors in $\complex{m}$.  Then
$\innerproduct{\vect{u}}{\vect{v}}=\conjugate{\innerproduct{\vect{v}}{\vect{u}}}$.</p>

</statement>

<proof>
<p>
<alignmath>
\innerproduct{\vect{u}}{\vect{v}}
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{v}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\,\conjugate{\conjugate{\vectorentry{\vect{v}}{i}}}]]>
<![CDATA[&&]]>\text{<acroref type="theorem" acro="CCT" />}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}\conjugate{\vectorentry{\vect{v}}{i}}}]]>
<![CDATA[&&]]>\text{<acroref type="theorem" acro="CCRM" />}\\
<![CDATA[&=\conjugate{\left(\sum_{i=1}^{m}\vectorentry{\vect{u}}{i}\conjugate{\vectorentry{\vect{v}}{i}}\right)}]]>
<![CDATA[&&]]>\text{<acroref type="theorem" acro="CCRA" />}\\
<![CDATA[&=\conjugate{\left(\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{v}}{i}}\vectorentry{\vect{u}}{i}\right)}]]>
<![CDATA[&&]]>\text{<acroref type="property" acro="CMCN" />}\\
<![CDATA[&=\conjugate{\innerproduct{\vect{v}}{\vect{u}}}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}\\
</alignmath>
</p>

</proof>
</theorem>

</subsection>

<subsection acro="N">
<title>Norm</title>

<p>If treating linear algebra in a more geometric fashion, the length of a vector occurs naturally, and is what you would expect from its name.   With complex numbers, we will define a similar function.  Recall that if $c$ is a complex number, then $\modulus{c}$ denotes its modulus (<acroref type="definition" acro="MCN" />).</p>

<definition acro="NV" index="vector!norm">
<title>Norm of a Vector</title>
<p>The <define>norm</define> of the vector $\vect{u}$ is the scalar quantity in $\complex{\null}$
<equation>
\norm{\vect{u}}=
\sqrt{
\modulus{\vectorentry{\vect{u}}{1}}^2+
\modulus{\vectorentry{\vect{u}}{2}}^2+
\modulus{\vectorentry{\vect{u}}{3}}^2+
\cdots+
\modulus{\vectorentry{\vect{u}}{m}}^2
}
=
\sqrt{\sum_{i=1}^{m}\modulus{\vectorentry{\vect{u}}{i}}^2}
</equation>
</p>

<notation acro="NV" index="norm">
<title>Norm of a Vector</title>
<usage>$\norm{\vect{v}}$</usage>
</notation>
</definition>

<p>Computing a norm is also easy to do.</p>

<example acro="CNSV" index="norm">
<title>Computing the norm of some vectors</title>

<p>The norm of
<equation>
\vect{u}=\colvector{3+2i\\1-6i\\2+4i\\2+i}
</equation>
is
<alignmath>
<![CDATA[\norm{\vect{u}}&=]]>
\sqrt{\modulus{3+2i}^2+\modulus{1-6i}^2+\modulus{2+4i}^2+\modulus{2+i}^2}\\
<![CDATA[&=\sqrt{13+37+20+5}=\sqrt{75}=5\sqrt{3}]]>
</alignmath>
</p>

<p>The norm of
<equation>
\vect{v}=\colvector{3\\-1\\2\\4\\-3}
</equation>
is
<equation>
\norm{\vect{v}}=
\sqrt{\modulus{3}^2+\modulus{-1}^2+\modulus{2}^2+\modulus{4}^2+\modulus{-3}^2}
=\sqrt{3^2+1^2+2^2+4^2+3^2}=\sqrt{39}.
</equation>
</p>

</example>

<p>Notice how the norm of a vector with real number entries is just the length of the vector.  Inner products and norms are related by the following theorem.</p>

<theorem acro="IPN" index="inner product!norm">
<title>Inner Products and Norms</title>
<statement>
<indexlocation index="norm!inner product" />
<p>Suppose that $\vect{u}$ is a vector in $\complex{m}$.  Then
$\norm{\vect{u}}^2=\innerproduct{\vect{u}}{\vect{u}}$.
</p>

</statement>

<proof>
<p>
<alignmath>
\norm{\vect{u}}^2
<![CDATA[&=\left(\sqrt{\sum_{i=1}^{m}\modulus{\vectorentry{\vect{u}}{i}}^2}\right)^2]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="NV" />}\\
<![CDATA[&=\sum_{i=1}^{m}\modulus{\vectorentry{\vect{u}}{i}}^2]]>
<![CDATA[&&]]>\text{Definition of square root}\\
<![CDATA[&=\sum_{i=1}^{m}\conjugate{\vectorentry{\vect{u}}{i}}\vectorentry{\vect{u}}{i}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="MCN" />}\\
<![CDATA[&=\innerproduct{\vect{u}}{\vect{u}}]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}
</alignmath>
</p>

</proof>
</theorem>

<p>When our vectors have entries only from the real numbers <acroref type="theorem" acro="IPN" /> says that the dot product of a vector with itself is equal to the length of the vector squared.</p>

<theorem acro="PIP" index="inner product!positive">
<title>Positive Inner Products</title>
<statement>
<p>Suppose that $\vect{u}$ is a vector in $\complex{m}$.  Then
$\innerproduct{\vect{u}}{\vect{u}}\geq 0$ with equality if and only if $\vect{u}=\zerovector$.</p>

</statement>

<proof>
<p>From the proof of <acroref type="theorem" acro="IPN" />  we see that
<equation>
\innerproduct{\vect{u}}{\vect{u}}
=
\modulus{\vectorentry{\vect{u}}{1}}^2+
\modulus{\vectorentry{\vect{u}}{2}}^2+
\modulus{\vectorentry{\vect{u}}{3}}^2+
\cdots+
\modulus{\vectorentry{\vect{u}}{m}}^2
</equation>
</p>

<p>Since each modulus is squared, every term is positive, and the sum must also be positive.  (Notice that in general the inner product is a complex number and cannot be compared with zero, but in the special case of $\innerproduct{\vect{u}}{\vect{u}}$ the result is a real number.)</p>

<p>The phrase, <q>with equality if and only if</q> means that we want to show that the statement $\innerproduct{\vect{u}}{\vect{u}}= 0$ (<ie /> with equality) is equivalent (<q>if and only if</q>) to the statement $\vect{u}=\zerovector$.</p>

<p>If $\vect{u}=\zerovector$, then it is a straightforward computation to see that $\innerproduct{\vect{u}}{\vect{u}}= 0$.  In the other direction, assume that $\innerproduct{\vect{u}}{\vect{u}}= 0$.  As before, $\innerproduct{\vect{u}}{\vect{u}}$ is a sum of moduli.  So we have
<equation>
0=\innerproduct{\vect{u}}{\vect{u}}=
\modulus{\vectorentry{\vect{u}}{1}}^2+
\modulus{\vectorentry{\vect{u}}{2}}^2+
\modulus{\vectorentry{\vect{u}}{3}}^2+
\cdots+
\modulus{\vectorentry{\vect{u}}{m}}^2
</equation></p>

<p>Now we have a sum of squares equaling zero, so each term must be zero.  Then by similar logic,
$\modulus{\vectorentry{\vect{u}}{i}}=0$
will imply that
$\vectorentry{\vect{u}}{i}=0$,
since $0+0i$ is the only complex number with zero modulus.  Thus every entry of $\vect{u}$ is zero and so $\vect{u}=\zerovector$, as desired.</p>

</proof>
</theorem>

<p>Notice that <acroref type="theorem" acro="PIP" /> contains <em>three</em> implications:
<alignmath>
<![CDATA[\vect{u}\in\complex{m}&\Rightarrow\innerproduct{\vect{u}}{\vect{u}}\geq 0\\]]>
<![CDATA[\vect{u}=\zerovector&\Rightarrow\innerproduct{\vect{u}}{\vect{u}}=0\\]]>
<![CDATA[\innerproduct{\vect{u}}{\vect{u}}=0&\Rightarrow\vect{u}=\zerovector]]>
</alignmath></p>

<p>The results contained in <acroref type="theorem" acro="PIP" /> are summarized by saying <q>the inner product is <define>positive definite</define>.</q></p>

<sageadvice acro="EVIC" index="exact vs.\ inexact computations">
<title>Exact Versus Inexact Computations</title>
We are now at a crossroads in our use of Sage.  So far our computations have involved rational numbers: fractions of two integers.  Sage is able to work with integers of seemingly unlimited size, and then can work with rational numbers exactly.  So all of our computations have been exactly correct so far.  In practice, many computations, especially those that originate with data, are not so precise.  Then we <em>represent</em> real numbers by <q>floating point numbers.</q>  Since the real numbers are infinite, finite computers must fake it with an extremely large, but still finite, collection of numbers.  The price we pay is that some computations will be just slightly imprecise when there is no number available that represents the exact answer.<br /><br />
You should now appreciate two problems that occur.  If we were to row-reduce a matrix with floating point numbers, there are potentially many computations and if a small amount of imprecision arises in each one, these errors can accumulate and lead to wildly incorrect answers.  When we row-reduce a matrix, whether or not an entry is zero or not is critically important in the decisions we make about which row operation to perform.  If we have an extremely small number (like $10^{-16}$) how can we be sure if it is zero or not?<br /><br />
Why discuss this now?  What is $\alpha=\sqrt{\frac{7}{3}}$?  Hard to say exactly, but it is definitely not a rational number.  Norms of vectors will feature prominently in all our discussions about orthogonal vectors, so we now have to recognize the need to work with square roots properly.  We have two strategies in Sage.<br /><br />
The number system <code>QQbar</code>, also known as the <q>field of algebraic numbers,</q> is a truly amazing feature of Sage.  It contains the rational numbers, plus <em>every</em> root of <em>every</em> polynomial with coefficients that are rational numbers.  For example, notice that $\alpha$ above is one solution to the polynomial equation $3x^2-7=0$ and thus is a number in <code>QQbar</code>, so Sage can work with it <em>exactly</em>.  These numbers are called <q>algebraic numbers</q> and you can recognize them since they print with a question mark near the end to remind you that when printed as a decimal they are approximations of numbers that Sage carries internally as exact quantities.  For example $\alpha$ can be created with <code>QQbar(sqrt(7/3))</code> and will print as <code>1.527525231651947</code>?.  Notice that complex numbers begin with the introduction of the imaginary number $i$, which is a root of the polynomial equation $x^2+1=0$, so the field of algebraic numbers contains many complex numbers.  The downside of <code>QQbar</code> is that computations are slow (relatively speaking), so this number system is most useful for examples and demonstrations.<br /><br />
The other strategy is to work strictly with approximate numbers, cognizant of the potential for inaccuracies.  Sage has two such number systems: <code>RDF</code> and <code>CDF</code>, which are comprised of <q>double precision</q> floating point numbers, first limited to just the reals, then expanded to the complexes.  Double-precision refers to the use of 64 bits to store the sign, mantissa and exponent in the representation of a real number.  This gives 53 bits of precision.  Do not confuse these fields with <code>RR</code> and <code>CC</code>, which are similar in appearance but very different in implementation.  Sage has implementations of several computations designed exclusively for <code>RDF</code> and <code>CDF</code>, such as the norm.  And they are very, very fast.  But some computations, like echelon form, can be wildly unreliable with these approximate numbers.  We will have more to say about this as we go.  In practice, you can use <code>CDF</code>, since <code>RDF</code> is a subset and only different in very limited cases.<br /><br />
In summary, <code>QQbar</code> is an extension of <code>QQ</code> which allows exact computations, but can be slow for large examples.  <code>RDF</code> and <code>CDF</code> are fast, with special algorithms to control much of the imprecision in some, but not all, computations.  So we need to be vigilant and skeptical when we work with these approximate numbers.  We will use both strategies, as appropriate.
</sageadvice>
<sageadvice acro="CNIP" index="conjugates, norms, inner products">
<title>Conjugates, Norms and Inner Products</title>
Conjugates, of complex numbers and of vectors, are straightforward, in <code>QQbar</code> or in <code>CDF</code>.
<sage>
<input>alpha = QQbar(2 + 3*I)
alpha.conjugate()
</input>
<output>2 - 3*I
</output>
</sage>

<sage>
<input>beta = CDF(2+3*I)
beta.conjugate()
</input>
<output>2.0 - 3.0*I
</output>
</sage>

<sage>
<input>v = vector(QQbar, [5-3*I, 2+6*I])
v.conjugate()
</input>
<output>(5 + 3*I, 2 - 6*I)
</output>
</sage>

<sage>
<input>w = vector(CDF, [5-3*I, 2+6*I])
w.conjugate()
</input>
<output>(5.0 + 3.0*I, 2.0 - 6.0*I)
</output>
</sage>

The term <q>inner product</q> means slightly different things to different people.  For some, it is the <q>dot product</q> that you may have seen in a calculus or physics course.  Our inner product could be called the <q>Hermitian inner product</q> to emphasize the use of vectors over the complex numbers and conjugating some of the entries.  So Sage has a <code>.dot_product()</code>, <code>.inner_product()</code>, and <code>.hermitian_inner_product()</code> <mdash /> we want to use the last one.<br /><br />
Furthermore, Sage defines the Hermitian inner product by conjugating entries from the <em>first</em> vector, rather than the <em>second</em> vector as defined in the text.  This is not as big a problem as it might seem.  First, <acroref type="theorem" acro="IPAC" />, tells us that we can counteract a difference in order by just taking the conjugate, so you can translate between Sage and the text by taking conjugates of results achieved from a Hermitian inner product.  Second, we will mostly be interested in when a Hermitian inner product is zero, which is its own conjugate, so no adjustment is required.  Third, most theorems are true as stated with either definition, although there are exceptions, like <acroref type="theorem" acro="IPSM" />.<br /><br />
From now on, when we mention an inner product in the context of using Sage, we will mean <code>.hermitian_inner_product()</code>.  We will redo the first part of <acroref type="example" acro="CSIP" />.  Notice that the syntax is a bit asymmetric.
<sage>
<input>u = vector(QQbar, [2+3*I,  5+2*I, -3+I])
v = vector(QQbar, [1+2*I, -4+5*I,  5*I])
u.hermitian_inner_product(v)
</input>
<output>3 + 19*I
</output>
</sage>

Again, notice that the result is the conjugate of what we have in the text.<br /><br />
Norms are as easy as conjugates.  Easier maybe.  It might be useful to realize that Sage uses entirely distinct code to compute an exact norm over <code>QQbar</code> versus an approximate norm over <code>CDF</code>, though that is totally transparent as you issue commands.  Here is <acroref type="example" acro="CNSV" /> reprised.


<sage>
<input>entries = [3+2*I, 1-6*I, 2+4*I, 2+I]
u = vector(QQbar, entries)
u.norm()
</input>
<output>8.66025403784439?
</output>
</sage>

<sage>
<input>u = vector(CDF, entries)
u.norm()
</input>
<output>8.66025403784
</output>
</sage>

<sage>
<input>numerical_approx(5*sqrt(3), digits = 30)
</input>
<output>8.66025403784438646763723170753
</output>
</sage>

We have three different numerical approximations, the latter 30-digit number being an approximation to the answer in the text.  But there is no inconsistency between them.  The first, an algebraic number, is represented internally as $5*a$ where $a$ is a root of the polynomial equation $x^2-3=0$, in other words it is $5\sqrt{3}$.  The <code>CDF</code> value prints with a few digits less than what is carried internally.  Notice that our different definitions of the inner product make no difference in the computation of a norm.<br /><br />
One warning now that we are working with complex numbers.  It is easy to <q>clobber</q> the symbol <code>I</code> used for the imaginary number $i$.  In other words, Sage will allow you to assign it to something else, rendering it useless.  An identity matrix is a likely reassignment.  If you run the next compute cell, be sure to evaluate the compute cell afterward to restore <code>I</code> to its usual role.
<sage>
<input>alpha = QQbar(5 - 6*I)
I = identity_matrix(2)
beta = QQbar(2+5*I)
</input>
<output>Traceback (most recent call last):
...
TypeError: Illegal initializer for algebraic number
</output>
</sage>

<sage>
<input>restore()
I^2
</input>
<output>-1
</output>
</sage>

We will finish with a verification of <acroref type="theorem" acro="IPN" />.  To test equality it is best if we work with entries from <code>QQbar</code>.
<sage>
<input>v = vector(QQbar, [2-3*I, 9+5*I, 6+2*I, 4-7*I])
v.hermitian_inner_product(v) == v.norm()^2
</input>
<output>True
</output>
</sage>

</sageadvice>
</subsection>

<subsection acro="OV">
<title>Orthogonal Vectors</title>

<p><q>Orthogonal</q> is a generalization of <q>perpendicular.</q>  You may have used mutually perpendicular vectors in a physics class, or you may recall from a calculus class that perpendicular vectors have a zero dot product.  We will now extend these ideas into the realm of higher dimensions and complex scalars.</p>

<definition acro="OV" index="orthogonal!vector pairs">
<title>Orthogonal Vectors</title>
<p>A pair of vectors, $\vect{u}$ and $\vect{v}$, from $\complex{m}$ are <define>orthogonal</define> if their inner product is zero, that is, $\innerproduct{\vect{u}}{\vect{v}}=0$.</p>

</definition>

<example acro="TOV" index="orthogonal vectors">
<title>Two orthogonal vectors</title>

<p>The vectors
<alignmath>
<![CDATA[\vect{u}&=\colvector{2 + 3i\\4 - 2i\\1 + i\\1 + i}]]>
<![CDATA[&]]>
<![CDATA[\vect{v}&=\colvector{1 - i\\2 + 3i\\4 - 6i\\1}]]>
</alignmath>
are orthogonal since
<alignmath>
\innerproduct{\vect{u}}{\vect{v}}
<![CDATA[&=(2-3i)(1-i)+(4+2i)(2+3i)+(1-i)(4-6i)+(1-i)(1)\\]]>
<![CDATA[&=(-1-5i)+(2+16i)+(-2-10i)+(1-i)\\]]>
<![CDATA[&=0+0i.]]>
</alignmath></p>

</example>

<p>We extend this definition to whole sets by requiring vectors to be pairwise orthogonal.  Despite using the same word, careful thought about what objects you are using will eliminate any source of confusion.</p>

<definition acro="OSV" index="orthogonal!set of vectors">
<title>Orthogonal Set of Vectors</title>
<p>Suppose that $S=\set{\vectorlist{u}{n}}$ is a set of vectors from $\complex{m}$.  Then $S$ is an <define>orthogonal set</define> if every pair of different vectors from $S$ is orthogonal, that is $\innerproduct{\vect{u}_i}{\vect{u}_j}=0$ whenever $i\neq j$.</p>

</definition>

<!-- TODO: Need a good example right here -->
<p>We now define the prototypical orthogonal set, which we will reference repeatedly.</p>

<definition acro="SUV" index="unit vectors">
<title>Standard Unit Vectors</title>
<p>Let $\vect{e}_j\in\complex{m}$, $1\leq j\leq m$ denote the column vectors defined by
<alignmath>
\vectorentry{\vect{e}_j}{i}
<![CDATA[&=]]>
\begin{cases}
<![CDATA[0&\text{if $i\neq j$}\\]]>
<![CDATA[1&\text{if $i=j$}]]>
\end{cases}
</alignmath>
</p>

<p>Then the set
<alignmath>
<![CDATA[\set{\vectorlist{e}{m}}&=\setparts{\vect{e}_j}{1\leq j\leq m}]]>
</alignmath>
is the set of <define>standard unit vectors</define> in $\complex{m}$.</p>

<notation acro="SUV" index="standard unit vector">
<title>Standard Unit Vectors</title>
<usage>$\vect{e}_i$</usage>
</notation>
</definition>

<p>Notice that $\vect{e}_j$ is identical to column $j$ of the $m\times m$ identity matrix $I_m$ (<acroref type="definition" acro="IM" />).  This observation will often be useful.  It is not hard to see that the set of standard unit vectors is an orthogonal set.  We will reserve the notation $\vect{e}_i$ for these vectors.</p>

<example acro="SUVOS" index="unit vectors!orthogonal">
<title>Standard Unit Vectors are an Orthogonal Set</title>

<p>Compute the inner product of two distinct vectors from the set of standard unit vectors (<acroref type="definition" acro="SUV" />), say $\vect{e}_i$, $\vect{e}_j$, where $i\neq j$,
<alignmath>
<![CDATA[\innerproduct{\vect{e}_i}{\vect{e}_j}&=]]>
\conjugate{0}0+
\conjugate{0}0+\cdots+
\conjugate{1}0+\cdots+
\conjugate{0}0+\cdots+
\conjugate{0}1+\cdots+
\conjugate{0}0+
\conjugate{0}0\\
<![CDATA[&=0(0)+0(0)+\cdots+1(0)+\cdots+0(1)+\cdots+0(0)+0(0)\\]]>
<![CDATA[&=0]]>
</alignmath></p>

<p>So the set $\set{\vectorlist{e}{m}}$ is an orthogonal set.</p>

</example>

<!--  Next example output (rescaled) from: -->
<!--  GramSchmidt[{{1+I,1, 1-I,I},{I,1+I,-1, -I}, {I,-I, -1+I,1}, {-1-I,I,1,-1}}, -->
<!--  InnerProduct(arrow)(Conjugate[#1].#2&)]//Simplify -->
<example acro="AOS" index="orthogonal!set">
<title>An orthogonal set</title>

<p>The set
<equation>
\set{\vect{x}_1,\,\vect{x}_2,\,\vect{x}_3,\,\vect{x}_4}=
\set{
\colvector{1+i\\1\\1-i\\i},\,
\colvector{1+5i\\6+5i\\-7-i\\1-6i},\,
\colvector{-7+34i\\-8-23i\\-10+22i\\30+13i},\,
\colvector{-2-4i\\6+i\\4+3i\\6-i}
}
</equation>
is an orthogonal set.</p>

<p>Since the inner product is anti-commutative (<acroref type="theorem" acro="IPAC" />) we can test pairs of different vectors in any order.  If the result is zero, then it will also be zero if the inner product is computed in the opposite order.  This means there are six different pairs of vectors to use in an inner product computation.  We'll do two and you can practice your inner products on the other four.
<alignmath>
<![CDATA[\innerproduct{\vect{x}_1}{\vect{x}_3}&=]]>
(1-i)(-7+34i)+(1)(-8-23i)+(1+i)(-10+22i)+(-i)(30+13i)\\
<![CDATA[&=(27+41i)+(-8-23i)+(-32+12i)+(13-30i)\\]]>
<![CDATA[&=0+0i]]>
<intertext>and</intertext>
<![CDATA[\innerproduct{\vect{x}_2}{\vect{x}_4}&=]]>
(1-5i)(-2-4i)+(6-5i)(6+i)+(-7+i)(4+3i)+(1+6i)(6-i)\\
<![CDATA[&=(-22+6i)+(41-24i)+(-31-17i)+(12+35i)\\]]>
<![CDATA[&=0+0i]]>
</alignmath></p>

</example>

<p>So far, this section has seen lots of definitions, and lots of theorems establishing un-surprising consequences of those definitions.  But here is our first theorem that suggests that inner products and orthogonal vectors have some utility.  It is also one of our first illustrations of how to arrive at linear independence as the conclusion of a theorem.</p>

<theorem acro="OSLI" index="orthogonal!linear independence">
<title>Orthogonal Sets are Linearly Independent</title>
<statement>
<indexlocation index="linear independence!orthogonal" />
<p>Suppose that $S$ is an orthogonal set of nonzero vectors.  Then $S$ is linearly independent.</p>

</statement>

<proof>
<p>Let $S=\set{\vectorlist{u}{n}}$ be an orthogonal set of nonzero vectors.  To prove the linear independence of $S$, we can appeal to the definition (<acroref type="definition" acro="LICV" />) and begin with an arbitrary relation of linear dependence (<acroref type="definition" acro="RLDCV" />),
<equation>
\lincombo{\alpha}{u}{n}=\zerovector.
</equation></p>

<p>Then, for every $1\leq i\leq n$, we have
<alignmath>
<![CDATA[&\alpha_i\innerproduct{\vect{u}_i}{\vect{u}_i}\\]]>
<![CDATA[&\quad\quad=\alpha_1(0)+\alpha_2(0)+\cdots+\alpha_i\innerproduct{\vect{u}_i}{\vect{u}_i}+\cdots+\alpha_n(0)]]>
<![CDATA[&&]]>\text{<acroref type="property" acro="ZCN" />}\\
<![CDATA[&\quad\quad=]]>
\alpha_1\innerproduct{\vect{u}_i}{\vect{u}_1}+
\cdots+
\alpha_i\innerproduct{\vect{u}_i}{\vect{u}_i}+
\cdots+
\alpha_n\innerproduct{\vect{u}_i}{\vect{u}_n}
<![CDATA[&&]]>\text{<acroref type="definition" acro="OSV" />}\\
<![CDATA[&\quad\quad=]]>
\innerproduct{\vect{u}_i}{\alpha_1\vect{u}_1}+
\innerproduct{\vect{u}_i}{\alpha_2\vect{u}_2}+
\cdots+
\innerproduct{\vect{u}_i}{\alpha_n\vect{u}_n}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPSM" />}\\
<![CDATA[&\quad\quad=]]>
\innerproduct{\vect{u}_i}{\lincombo{\alpha}{u}{n}}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPVA" />}\\
<![CDATA[&\quad\quad=]]>
\innerproduct{\vect{u}_i}{\zerovector}
<![CDATA[&&]]>\text{<acroref type="definition" acro="RLDCV" />}\\
<![CDATA[&\quad\quad=0]]>
<![CDATA[&&]]>\text{<acroref type="definition" acro="IP" />}
</alignmath></p>

<p>Because $\vect{u}_i$ was assumed to be nozero, <acroref type="theorem" acro="PIP" /> says $\innerproduct{\vect{u}_i}{\vect{u}_i}$ is nonzero and thus $\alpha_i$ must be zero.  So we conclude that $\alpha_i=0$ for all $1\leq i\leq n$ in any relation of linear dependence on $S$.  But this says that $S$ is a linearly independent set since the only way to form a relation of linear dependence is the trivial way (<acroref type="definition" acro="LICV" />).  Boom!</p>

</proof>
</theorem>

</subsection>

<subsection acro="GSP">
<title>Gram-Schmidt Procedure</title>

<p>The Gram-Schmidt Procedure is really a theorem.  It says that if we begin with a linearly independent set of $p$ vectors, $S$, then we can do a number of calculations with these vectors and produce an orthogonal set of $p$ vectors, $T$, so that $\spn{S}=\spn{T}$.  Given the large number of computations involved, it is indeed a procedure to do all the necessary computations, and it is best employed on a computer.  However, it also has value in proofs where we may on occasion wish to replace a linearly independent set by an orthogonal set.</p>

<p>This is our first occasion to use the technique of <q>mathematical induction</q> for a proof, a technique we will see again several times, especially in <acroref type="chapter" acro="D" />.  So study the simple example described in <acroref type="technique" acro="I" /> first.</p>

<theorem acro="GSP" index="Gram-Schmidt!column vectors">
<title>Gram-Schmidt Procedure</title>
<statement>
<p>Suppose that $S=\set{\vectorlist{v}{p}}$ is a linearly independent set of vectors in $\complex{m}$.  Define the vectors $\vect{u}_i$, $1\leq i\leq p$ by
<equation>
\vect{u}_i=\vect{v}_i
-\frac{\innerproduct{\vect{u}_1}{\vect{v}_i}}{\innerproduct{\vect{u}_1}{\vect{u}_1}}\vect{u}_1
-\frac{\innerproduct{\vect{u}_2}{\vect{v}_i}}{\innerproduct{\vect{u}_2}{\vect{u}_2}}\vect{u}_2
-\frac{\innerproduct{\vect{u}_3}{\vect{v}_i}}{\innerproduct{\vect{u}_3}{\vect{u}_3}}\vect{u}_3
-\cdots
-\frac{\innerproduct{\vect{u}_{i-1}}{\vect{v}_i}}{\innerproduct{\vect{u}_{i-1}}{\vect{u}_{i-1}}}\vect{u}_{i-1}
</equation></p>

<p>Then if $T=\set{\vectorlist{u}{p}}$, then $T$ is an orthogonal set of non-zero vectors, and $\spn{T}=\spn{S}$.</p>

</statement>

<proof>
<p>We will prove the result by using induction on $p$ (<acroref type="technique" acro="I" />).  To begin, we prove that $T$ has the desired properties when $p=1$.  In this case $\vect{u}_1=\vect{v}_1$ and $T=\set{\vect{u}_1}=\set{\vect{v}_1}=S$.  Because $S$ and $T$ are equal, $\spn{S}=\spn{T}$.  Equally trivial, $T$ is an orthogonal set.  If $\vect{u}_1=\zerovector$, then $S$ would be a linearly dependent set, a contradiction.</p>

<p>Suppose that the theorem is true for any set of $p-1$ linearly independent vectors.  Let $S=\set{\vectorlist{v}{p}}$ be a linearly independent set of $p$ vectors.  Then $S^\prime=\set{\vectorlist{v}{p-1}}$ is also linearly independent.  So we can apply the theorem to $S^\prime$ and construct the vectors $T^\prime=\set{\vectorlist{u}{p-1}}$.  $T^\prime$ is therefore an orthogonal set of nonzero vectors and $\spn{S^\prime}=\spn{T^\prime}$.  Define
<equation>
\vect{u}_p=\vect{v}_p
-\frac{\innerproduct{\vect{u}_1}{\vect{v}_p}}{\innerproduct{\vect{u}_1}{\vect{u}_1}}\vect{u}_1
-\frac{\innerproduct{\vect{u}_2}{\vect{v}_p}}{\innerproduct{\vect{u}_2}{\vect{u}_2}}\vect{u}_2
-\frac{\innerproduct{\vect{u}_3}{\vect{v}_p}}{\innerproduct{\vect{u}_3}{\vect{u}_3}}\vect{u}_3
-\cdots
-\frac{\innerproduct{\vect{u}_{p-1}}{\vect{v}_p}}{\innerproduct{\vect{u}_{p-1}}{\vect{u}_{p-1}}}\vect{u}_{p-1}
</equation>
and let $T=T^\prime\cup\set{\vect{u}_p}$.  We need to now show that $T$ has several properties by building on what we know about $T^\prime$.  But first notice that the above equation has no problems with the denominators ($\innerproduct{\vect{u}_i}{\vect{u}_i}$) being zero, since the $\vect{u}_i$ are from $T^\prime$, which is composed of nonzero vectors.</p>

<p>We show that $\spn{T}=\spn{S}$, by first establishing that $\spn{T}\subseteq\spn{S}$.  Suppose $\vect{x}\in\spn{T}$, so
<equation>
\vect{x}=\lincombo{a}{u}{p}
</equation>
The term $a_p\vect{u}_p$ is a linear combination of vectors from $T^\prime$ and the vector $\vect{v}_p$, while the remaining terms are a linear combination of vectors from $T^\prime$.  Since $\spn{T^\prime}=\spn{S^\prime}$, any term that is a multiple of a vector from $T^\prime$ can be rewritten as a linear combination of vectors from $S^\prime$.  The remaining term $a_p\vect{v}_p$ is a multiple of a vector in $S$.  So we see that $\vect{x}$ can be rewritten as a linear combination of vectors from $S$, <ie /> $\vect{x}\in\spn{S}$.</p>

<p>To show that $\spn{S}\subseteq\spn{T}$, begin with $\vect{y}\in\spn{S}$, so
<equation>
\vect{y}=\lincombo{a}{v}{p}
</equation>
</p>

<p>Rearrange our defining equation for $\vect{u}_p$ by solving for $\vect{v}_p$.  Then the term $a_p\vect{v}_p$ is a multiple of a linear combination of elements of $T$.  The remaining terms are a linear combination of $\vectorlist{v}{p-1}$, hence an element of $\spn{S^\prime}=\spn{T^\prime}$.  Thus these remaining terms can be written as a linear combination of the vectors in $T^\prime$.   So $\vect{y}$ is a linear combination of vectors from $T$, <ie /> $\vect{y}\in\spn{T}$.</p>

<p>The elements of $T^\prime$ are nonzero, but what about $\vect{u}_p$?  Suppose to the contrary that $\vect{u}_p=\zerovector$,
<alignmath>
<![CDATA[\zerovector&=\vect{u}_p=\vect{v}_p]]>
-\frac{\innerproduct{\vect{u}_1}{\vect{v}_p}}{\innerproduct{\vect{u}_1}{\vect{u}_1}}\vect{u}_1
-\frac{\innerproduct{\vect{u}_2}{\vect{v}_p}}{\innerproduct{\vect{u}_2}{\vect{u}_2}}\vect{u}_2
-\frac{\innerproduct{\vect{u}_3}{\vect{v}_p}}{\innerproduct{\vect{u}_3}{\vect{u}_3}}\vect{u}_3
-\cdots
-\frac{\innerproduct{\vect{u}_{p-1}}{\vect{v}_p}}{\innerproduct{\vect{u}_{p-1}}{\vect{u}_{p-1}}}\vect{u}_{p-1}\\
<![CDATA[&\vect{v}_p=]]>
 \frac{\innerproduct{\vect{u}_1}{\vect{v}_p}}{\innerproduct{\vect{u}_1}{\vect{u}_1}}\vect{u}_1
+\frac{\innerproduct{\vect{u}_2}{\vect{v}_p}}{\innerproduct{\vect{u}_2}{\vect{u}_2}}\vect{u}_2
+\frac{\innerproduct{\vect{u}_3}{\vect{v}_p}}{\innerproduct{\vect{u}_3}{\vect{u}_3}}\vect{u}_3
+\cdots
+\frac{\innerproduct{\vect{u}_{p-1}}{\vect{v}_p}}{\innerproduct{\vect{u}_{p-1}}{\vect{u}_{p-1}}}\vect{u}_{p-1}
</alignmath></p>

<p>Since $\spn{S^\prime}=\spn{T^\prime}$ we can write the vectors $\vectorlist{u}{p-1}$ on the right side of this equation in terms of the vectors $\vectorlist{v}{p-1}$ and we then have the vector $\vect{v}_p$ expressed as a linear combination of the other $p-1$ vectors in $S$, implying that $S$ is a linearly dependent set (<acroref type="theorem" acro="DLDS" />), contrary to our lone hypothesis about $S$.</p>

<p>Finally, it is a simple matter to establish that $T$ is an orthogonal set, though it will not appear so simple looking.  Think about your objects as you work through the following <mdash /> what is a vector and what is a scalar.  Since $T^\prime$ is an orthogonal set by induction, most pairs of elements in $T$ are already known to be orthogonal.  We just need to test <q>new</q> inner products, between $\vect{u}_p$ and $\vect{u}_i$, for $1\leq i\leq p-1$.  Here we go, using summation notation,
<alignmath>
<![CDATA[\innerproduct{\vect{u}_i}{\vect{u}_p}&=]]>
\innerproduct{\vect{u}_i}{
\vect{v}_p-\sum_{k=1}^{p-1}\frac{\innerproduct{\vect{u}_k}{\vect{v}_p}}{\innerproduct{\vect{u}_k}{\vect{u}_k}}\vect{u}_k
}
\\
<![CDATA[&=]]>
\innerproduct{\vect{u}_i}{\vect{v}_p}
-
\innerproduct{\vect{u}_i}{
\sum_{k=1}^{p-1}\frac{\innerproduct{\vect{u}_k}{\vect{v}_p}}{\innerproduct{\vect{u}_k}{\vect{u}_k}}\vect{u}_k
}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPVA" />}\\
<![CDATA[&=]]>
\innerproduct{\vect{u}_i}{\vect{v}_p}
-
\sum_{k=1}^{p-1}\innerproduct{\vect{u}_i}{
\frac{\innerproduct{\vect{u}_k}{\vect{v}_p}}{\innerproduct{\vect{u}_k}{\vect{u}_k}}\vect{u}_k
}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPVA" />}\\
<![CDATA[&=]]>
\innerproduct{\vect{u}_i}{\vect{v}_p}
-
\sum_{k=1}^{p-1}\frac{\innerproduct{\vect{u}_k}{\vect{v}_p}}{\innerproduct{\vect{u}_k}{\vect{u}_k}}\innerproduct{\vect{u}_i}{\vect{u}_k}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPSM" />}\\
<![CDATA[&=]]>
\innerproduct{\vect{u}_i}{\vect{v}_p}
-
\frac{\innerproduct{\vect{u}_i}{\vect{v}_p}}{\innerproduct{\vect{u}_i}{\vect{u}_i}}\innerproduct{\vect{u}_i}{\vect{u}_i}
-
\sum_{k\neq i}\frac{\innerproduct{\vect{u}_k}{\vect{v}_p}}{\innerproduct{\vect{u}_k}{\vect{u}_k}}(0)
<![CDATA[&&\text{Induction Hypothesis}\\]]>
<![CDATA[&=]]>
\innerproduct{\vect{u}_i}{\vect{v}_p}
-
\innerproduct{\vect{u}_i}{\vect{v}_p}
-
\sum_{k\neq i}0\\
<![CDATA[&=0]]>
</alignmath>
</p>

</proof>
</theorem>

<example acro="GSTV" index="Gram-Schmidt!three vectors">
<title>Gram-Schmidt of three vectors</title>

<p>We will illustrate the Gram-Schmidt process with three vectors.  Begin with the linearly independent (check this!) set
<equation>
S=\set{\vect{v}_1,\,\vect{v}_2,\,\vect{v}_3}=\set{
\colvector{1\\1+i\\1},\,
\colvector{-i\\1\\1+i},\,
\colvector{0\\i\\i}
}
</equation></p>

<p>Then
<alignmath>
<![CDATA[\vect{u}_1&=\vect{v_1}=\colvector{1\\1+i\\1}\\]]>
<![CDATA[\vect{u}_2&=\vect{v}_2]]>
-\frac{\innerproduct{\vect{u}_1}{\vect{v}_2}}{\innerproduct{\vect{u}_1}{\vect{u}_1}}\vect{u}_1
=\frac{1}{4}\colvector{-2-3i\\1-i\\2+5i}\\
<![CDATA[\vect{u}_3&=\vect{v}_3]]>
-\frac{\innerproduct{\vect{u}_1}{\vect{v}_3}}{\innerproduct{\vect{u}_1}{\vect{u}_1}}\vect{u}_1
-\frac{\innerproduct{\vect{u}_2}{\vect{v}_3}}{\innerproduct{\vect{u}_2}{\vect{u}_2}}\vect{u}_2
=\frac{1}{11}\colvector{-3-i\\1+3i\\-1-i}
</alignmath>
and
<equation>
T=\set{\vect{u}_1,\,\vect{u}_2,\,\vect{u}_3}
=\set{
\colvector{1\\1+i\\1},\,
\frac{1}{4}\colvector{-2-3i\\1-i\\2+5i},\,
\frac{1}{11}\colvector{-3-i\\1+3i\\-1-i}
}
</equation>
is an orthogonal set (which you can check) of nonzero vectors and $\spn{T}=\spn{S}$ (all by <acroref type="theorem" acro="GSP" />).  Of course, as a by-product of orthogonality, the set $T$ is also linearly independent (<acroref type="theorem" acro="OSLI" />).</p>

</example>

<p>One final definition related to orthogonal vectors.</p>

<definition acro="ONS" index="orthonormal">
<title>OrthoNormal Set</title>
<p>Suppose $S=\set{\vectorlist{u}{n}}$ is an orthogonal set of vectors such that $\norm{\vect{u}_i}=1$ for all $1\leq i\leq n$.  Then $S$ is an <define>orthonormal</define> set of vectors.</p>

</definition>

<p>Once you have an orthogonal set, it is easy to convert it to an orthonormal set <mdash /> multiply each vector by the reciprocal of its norm, and the resulting vector will have norm 1.  This scaling of each vector will not affect the orthogonality properties (apply <acroref type="theorem" acro="IPSM" />).</p>

<example acro="ONTV" index="orthonormal set!three vectors">
<title>Orthonormal set, three vectors</title>

<p>The set
<equation>
T=\set{\vect{u}_1,\,\vect{u}_2,\,\vect{u}_3}
=\set{
\colvector{1\\1+i\\1},\,
\frac{1}{4}\colvector{-2-3i\\1-i\\2+5i},\,
\frac{1}{11}\colvector{-3-i\\1+3i\\-1-i}
}
</equation>
from <acroref type="example" acro="GSTV" /> is an orthogonal set.</p>

<p>We compute the norm of each vector,
<alignmath>
<![CDATA[\norm{\vect{u}_1}=2&&]]>
<![CDATA[\norm{\vect{u}_2}=\frac{1}{2}\sqrt{11}&&]]>
\norm{\vect{u}_3}=\frac{\sqrt{2}}{\sqrt{11}}
</alignmath>
</p>

<p>Converting each vector to a norm of 1, yields an orthonormal set,
<alignmath>
<![CDATA[\vect{w}_1&=\frac{1}{2}\colvector{1\\1+i\\1}\\]]>
<![CDATA[\vect{w}_2&=\frac{1}{\frac{1}{2}\sqrt{11}}\frac{1}{4}\colvector{-2-3i\\1-i\\2+5i}=\frac{1}{2\sqrt{11}}\colvector{-2-3i\\1-i\\2+5i}\\]]>
<![CDATA[\vect{w}_3&=\frac{1}{\frac{\sqrt{2}}{\sqrt{11}}}\frac{1}{11}\colvector{-3-i\\1+3i\\-1-i}=\frac{1}{\sqrt{22}}\colvector{-3-i\\1+3i\\-1-i}]]>
</alignmath>
</p>

</example>

<example acro="ONFV" index="orthonormal set!four vectors">
<title>Orthonormal set, four vectors</title>

<p>As an exercise convert the linearly independent set
<equation>
S=\set{
\colvector{1+i\\1\\1-i\\i},\,
\colvector{i\\1+i\\-1\\-i},\,
\colvector{i\\-i\\ -1+i\\1},\,
\colvector{-1-i\\i\\1\\-1}
}
</equation>
to an orthogonal set via the Gram-Schmidt Process (<acroref type="theorem" acro="GSP" />) and then scale the vectors to norm 1 to create an orthonormal set.  You should get the same set you would if you scaled the orthogonal set of <acroref type="example" acro="AOS" /> to become an orthonormal set.</p>

</example>

<p>We will see orthonormal sets again in <acroref type="subsection" acro="MINM.UM" />.   They are intimately related to unitary matrices (<acroref type="definition" acro="UM" />) through <acroref type="theorem" acro="CUMOS" />.  Some of the utility of orthonormal sets is captured by <acroref type="theorem" acro="COB" /> in <acroref type="subsection" acro="B.OBC" />.   Orthonormal sets appear once again in <acroref type="section" acro="OD" /> where they are key in orthonormal diagonalization.</p>

<sageadvice acro="OGS" index="orthogonality and Gram-Schmidt">
<title>Orthogonality and Gram-Schmidt</title>
It is easy enough to check a pair of vectors for orthogonality (is the inner product zero?).  To check that a set is orthogonal, we just need to do this repeatedly. This is a redo of <acroref type="example" acro="AOS" />.
<sage>
<input>x1 = vector(QQbar, [    1+I,       1,      1-I,       I])
x2 = vector(QQbar, [  1+5*I,   6+5*I,     -7-I,   1-6*I])
x3 = vector(QQbar, [-7+34*I, -8-23*I, -10+22*I, 30+13*I])
x4 = vector(QQbar, [ -2-4*I,     6+I,    4+3*I,     6-I])
S = [x1, x2, x3, x4]
ips = [S[i].hermitian_inner_product(S[j])
                  for i in range(3) for j in range(i+1,3)]
all([ip == 0 for ip in ips])
</input>
<output>True
</output>
</sage>

Notice how the list comprehension computes each pair just once, and never checks the inner product of a vector with itself.  If we wanted to check that a set is orthonormal, the <q>normal</q> part is less involved.  We will check the set above, even though we can clearly see that the four vectors are not even close to being unit vectors.  Be sure to run the above definitions of <code>S</code> before running the next compute cell.
<sage>
<input>ips = [S[i].hermitian_inner_product(S[i]) for i in range(3)]
all([ip == 1 for ip in ips])
</input>
<output>False
</output>
</sage>

Applying the Gram-Schmidt procedure to a set of vectors is the type of computation that a program like Sage is perfect for.  Gram-Schmidt is implemented as a method for matrices, where we interpret the rows of the matrix as the vectors in the original set.  The result is two matrices, where the first has rows that are the orthogonal vectors.  The second matrix has rows that provide linear combinations of the orthogonal vectors that equal the original vectors.  The original vectors do not need to form a linearly independent set, and when the set is linearly dependent, then zero vectors produced are not part of the returned set.<br /><br />
Over <code>CDF</code> the set is automatically orthonormal, and since a different algorithm is used (to help control the imprecisions), the results will look different than what would result from <acroref type="theorem" acro="GSP" />.  We will illustrate with the vectors from <acroref type="example" acro="GSTV" />.
<sage>
<input>v1 = vector(CDF, [ 1, 1+I,   1])
v2 = vector(CDF, [-I,   1, 1+I])
v3 = vector(CDF, [ 0,   I,   I])
A = matrix([v1,v2,v3])
G, M = A.gram_schmidt()
G.round(5)
</input>
<output>[                -0.5         -0.5 - 0.5*I                 -0.5]
[ 0.30151 + 0.45227*I -0.15076 + 0.15076*I -0.30151 - 0.75378*I]
[   0.6396 + 0.2132*I   -0.2132 - 0.6396*I    0.2132 + 0.2132*I]
</output>
</sage>

We formed the matrix A with the three vectors as rows, and of the two outputs we are interested in the first one, whose rows form the orthonormal set.  We round the numbers to 5 digits, just to make the result fit nicely on your screen.  Let's do it again, now exactly over <code>QQbar</code>.  We will output the entries of the matrix as list, working across rows first, so it fits nicely.
<sage>
<input>v1 = vector(QQbar, [ 1, 1+I,   1])
v2 = vector(QQbar, [-I,   1, 1+I])
v3 = vector(QQbar, [ 0,   I,   I])
A = matrix([v1,v2,v3])
G, M = A.gram_schmidt(orthonormal=True)
Sequence(G.list(), cr=True)
</input>
<output>[
0.50000000000000000?,
0.50000000000000000? + 0.50000000000000000?*I,
0.50000000000000000?,
-0.3015113445777636? - 0.4522670168666454?*I,
0.1507556722888818? - 0.1507556722888818?*I,
0.3015113445777636? + 0.7537783614444091?*I,
-0.6396021490668313? - 0.2132007163556105?*I,
0.2132007163556105? + 0.6396021490668313?*I,
-0.2132007163556105? - 0.2132007163556105?*I
]
</output>
</sage>

Notice that we asked for orthonormal output, so the rows of <code>G</code> are the vectors $\set{\vect{w}_1,\,\vect{w}_2,\,\vect{w}_3}$ in <acroref type="example" acro="ONTV" />.  Exactly.  We can restrict ourselves to <code>QQ</code> and forego the <q>normality</q> to obtain just the orthogonal set $\set{\vect{u}_1,\,\vect{u}_2,\,\vect{u}_3}$ of <acroref type="example" acro="GSTV" />.
<sage>
<input>v1 = vector(QQbar, [ 1, 1+I,   1])
v2 = vector(QQbar, [-I,   1, 1+I])
v3 = vector(QQbar, [ 0,   I,   I])
A = matrix([v1, v2, v3])
G, M = A.gram_schmidt(orthonormal=False)
Sequence(G.list(), cr=True)
</input>
<output>[
1,
I + 1,
1,
-0.50000000000000000? - 0.75000000000000000?*I,
0.25000000000000000? - 0.25000000000000000?*I,
0.50000000000000000? + 1.2500000000000000?*I,
-0.2727272727272728? - 0.0909090909090909?*I,
0.0909090909090909? + 0.2727272727272728?*I,
-0.0909090909090909? - 0.0909090909090909?*I
]
</output>
</sage>

Notice that it is an error to ask for an orthonormal set over <code>QQ</code> since you cannot expect to take square roots of rationals and stick with rationals.
<sage>
<input>v1 = vector(QQ, [1, 1])
v2 = vector(QQ, [2, 3])
A = matrix([v1,v2])
G, M = A.gram_schmidt(orthonormal=True)
</input>
<output>Traceback (most recent call last):
...
TypeError: QR decomposition unable to compute square roots in Rational Field
</output>
</sage>



</sageadvice>
</subsection>

<!--   End  o.tex -->
<readingquestions>
<ol>
<li>Is the set
<equation>
\set{\colvector{1\\-1\\2},\,\colvector{5\\3\\-1},\,\colvector{8\\4\\-2}}
</equation>
an orthogonal set?  Why?
</li>
<li>What is the distinction between an orthogonal set and an orthonormal set?
</li>
<li> What is nice about the output of the Gram-Schmidt process?
</li></ol>
</readingquestions>

<exercisesubsection>

<exercise type="C" number="20" rough="4 orthogonalities in Example AOS">
<problem contributor="robertbeezer">Complete <acroref type="example" acro="AOS" /> by verifying that the four remaining inner products are zero.
</problem>
</exercise>

<exercise type="C" number="21" rough="Example GSTV, verify GS output">
<problem contributor="robertbeezer">Verify that the set $T$ created in <acroref type="example" acro="GSTV" /> by the Gram-Schmidt Procedure is an orthogonal set.
</problem>
</exercise>

<exercise type="M" number="60" rough="{u,v,w} on, then u+v, v+w not orthogonal">
<problem contributor="manleyperkel">Suppose that $\set{\vect{u},\,\vect{v},\,\vect{w}}\subseteq\complex{n}$ is an orthonormal set.  Prove that $\vect{u}+\vect{v}$ is not orthogonal to $\vect{v}+\vect{w}$.
</problem>
</exercise>

<exercise type="T" number="10" rough="Prove part 1 Theorem IPVA">
<problem contributor="robertbeezer">Prove part 2 of the conclusion of <acroref type="theorem" acro="IPVA" />.
</problem>
</exercise>

<exercise type="T" number="11" rough="Prove part 1 Theorem IPSM">
<problem contributor="robertbeezer">Prove part 2 of the conclusion of <acroref type="theorem" acro="IPSM" />.
</problem>
</exercise>

<exercise type="T" number="20" rough="Orthogonal to linear combo">
<problem contributor="robertbeezer">Suppose that $\vect{u},\,\vect{v},\,\vect{w}\in\complex{n}$, $\alpha,\,\beta\in\complexes$ and $\vect{u}$ is orthogonal to both $\vect{v}$ and $\vect{w}$.  Prove that $\vect{u}$ is orthogonal to $\alpha\vect{v}+\beta\vect{w}$.
</problem>
<solution contributor="robertbeezer">Vectors are orthogonal if their inner product is zero (<acroref type="definition" acro="OV" />), so we compute,
<alignmath>
\innerproduct{\vect{u}}{\alpha\vect{v}+\beta\vect{w}}
<![CDATA[&=]]>
\innerproduct{\vect{u}}{\alpha\vect{v}}+
\innerproduct{\vect{u}}{\beta\vect{w}}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPVA" />}\\
<![CDATA[&=]]>
\alpha\innerproduct{\vect{u}}{\vect{v}}+
\beta\innerproduct{\vect{u}}{\vect{w}}
<![CDATA[&&]]>\text{<acroref type="theorem" acro="IPSM" />}\\
<![CDATA[&=]]>
\alpha\left(0\right)+\beta\left(0\right)
<![CDATA[&&]]>\text{<acroref type="definition" acro="OV" />}\\
<![CDATA[&=0]]>
</alignmath>
So by <acroref type="definition" acro="OV" />, $\vect{u}$ and $\alpha\vect{v}+\beta\vect{w}$ are an orthogonal pair of vectors.
</solution>
</exercise>

<exercise type="T" number="30" rough="Pointless to Gram-Schmidt an orthogonal set">
<problem contributor="stevecanfield">Suppose that the set $S$ in the hypothesis of <acroref type="theorem" acro="GSP" /> is not just linearly independent, but is also orthogonal.  Prove that the set $T$ created by the Gram-Schmidt procedure is equal to $S$.  (Note that we are getting a stronger conclusion than $\spn{T}=\spn{S}$ <mdash /> the conclusion is that $T=S$.)  In other words, it is pointless to apply the Gram-Schmidt procedure to a set that is already orthogonal.
</problem>
</exercise>

</exercisesubsection>

</section>