1第1章信息论基础1.736136236336436536636536436336236112111098765432)(XqX1.8p(s0)=0.8p(s0)+0.5p(s2)p(s1)=0.2p(s0)+0.5p(s2)p(s2)=0.5p(s1)+0.3p(s3)p(s3)=0.5p(s1)+0.7p(s3)p(s0)+p(s1)+p(s2)+p(s3)=1p(s0)=3715,p(s1)=p(s2)=376,p(s3)=37101.9Pe=q(0)p+q(1)p=0.06(1-0.06)﹡1000﹡10=94009500不能1.1022222222)1(0)1()1(00000)1(0)1()1(000000)1()1(0)1(00000)1()1(0)1(ppppppppppppppppppppppppP第2章信息的度量2.4logk2.5I(X;YZ)=I(X;Y)+I(X;Z∣Y)2.7010434()()()111111pspspsH=0.25(Bit/符号)2.8H=0.82(Bit/符号)2.10(1)1()log225.6()52!iIxBit(2)1352!()log()log413!39!iiIxqx(3))/(4.713log234log52log521log)(符号-BitUH(4))/(7.313log131log)(符号BitXH2.11(1)H(X)=log6=2.58(Bit/符号)(2)H(X)=2.36(Bit/符号)(3)I(A+B=7)=-log1/6=log6=2.585(Bit)2.12(1)I(xi)=-log1/100=log100(2)H(X)=log100.22.13039.0log)(YXI2.14Rt=1000/4(码字/秒)×H(U)=250×9=2250(Bit/秒)2.15―logp=log55/44。2.16I(X;Y)=1.82(Bit/符号)2.17(1))1(2log100;14puI(2))1(2log2100;104puI(3))1(2log3100;1004puI2.20kBkkAAACeq111时,H(U)取得最大值。0110)1(log)1(1log)(kkkkkkkmAAAAkqqUH2.30(1)H(X)=1.69(Bit/符号)(2)H(Y)=1.57(Bit/符号)(3)符号/76.0)(BitXYH(4)(Bit/2.45)(符号XYH(5)I(X;Y)=0.81(Bit/符号)第3章离散信源无失真编码3.6(1)2位(2)取码长n1=1、n1=2、n1=3、n1=3就能得到紧致码。3.13(1)H(X)=-0.9log0.9-0.1log0.1=log10-0.9log9(2)7.5l(3)7.2n3.14(1)x1→0,x2→11,x3→105.1n99.0(2)x1x1→01,x1x2→110,x2x1→101,x1x3→011,x3x1→010,x2x2→1111,x2x3→1110,x3x2→10013n99.03.15(1))(112pHpH3(2)pn113.16(1)当M=2i,这种情况下得到的是等长码,码长为i。(2)当M=2i+1平均码长为122)1(1221212iiiiiiin3.17方法一:概率之和与原信源某概率相等,概率之和往上排:方法二:概率之和与原信源某概率相等,概率之和往下排:两种方法得到的码集都是最佳的。3.18(1)shannon编码,D=2消息x1x2x3x4x5x6x7x8码字00000101110001010101111011110(2)Fano编码,D=2消息x1x2x3x4x5x6x7x8码字0001001110010111011101111(3)Huffman编码,D=3消息x1x2x3x4x5x6x7x8码字02120121110221220(4)Huffman编码,D=4消息x1x2x3x4x5x6x7x8码字203332313011103.21(1)D=2消息x1x2x3x4x5x6x7x8x9x10码字1111011000110010001101110001010100(2)D=3消息x1x2x3x4x5x6x7x8x9x10码字22212012100201001111103.22方法一:概率之和与原信源某概率相等,概率之和往上排;方法二:概率之和与原信源某概率相等,概率之和往下排;第一种方法对实用更好3.24(1)H(X)=(1/4)log4+(3/4)log(4/3)=0.811966.2n984.0467.3n842.0967.1n937.0633.1n89.026.3n99.011.2n966.04(2)q(0)=1/4,q(1)=3/4(3)扩展信源16/916/316/316/111011000xxxxxxxxFano编码消息x0x0x0x1x1x0x1x1码字111110100(4)扩展信源64/2764/964/964/364/964/364/364/1111011101001110010100000xxxxxxxxxxxxxxxxxxxxxxxxHuffman编码D=2消息x1x1x1x1x1x0x1x0x1x0x1x1x1x0x0x0x1x0x0x0x1x0x0x0码字0101101100111111111011101111003.25(1)Shannon编码D=2消息x1x2x3x4码字010110111(2)Huffman编码D=2,得到的编码与(1)一致(3)之所以得到这种结果,是因为消息的每个概率都可以写成D–ni的形式,取每个码字的码长分别为ni即可。第4章离散信道的信道容量4.7(1))(3log2pHC(2))(2log2pHC(3))/(322.0符号BitC(4)C=e2-(3/2)(5))/(083.0符号BitC(6))/(082.0符号BitC4.8(1))/(918.2符号BitC(2)2.0log2.05.0log5.035.0log7.0C4.9(1))()1();(2HqYXI(2)α=0.5时,I(X;Y)取得最大值2log)1(qC96.06875.1n985.047.2n5(3)2log)0;0(I0);0()0;1(eII4.11符号)(/0312.0BitC4.12(1)log)1log()1()1()1log()1(---C(2)2/)1(1logI(1;1)2/)1(logI(1;0)4.14(1))(221lnHeC(2)3ln]21ln[21ln0)(0)(2eeCH3ln]21ln[21ln1)(0)(2eeCH)2ln21ln0.5)(eC4.15C=0.322(Bit/码符)4.16(1)I(X;Y)=1.5(Bit/符号)(2)I(X;Z)=1.5(Bit/符号)(3)I(X;Z)和I(X;Y)二者相等第5章习题5.10(1)采用极大后验概率译码准则:y1→x1,y2→x2,y3→x1pe1=11/24(2)信源等概分布,采用极大似然译码准则:y1→x1,y2→x2,y3→x3pe1=1/25.11译码:y1→x1,y2→x3,y3→x2pe1=21/425.12(1)编码x0:00,x1:11,x2:22,x3:33,x4:44译码00,01,10→00,11,12,21→11,22,23,32→22,33,34,43→33,40,44,04→44pe=1/4(2)编码:x0:02,x1:14,x2:21,x3:33,x4:40根据最大似然函数译码准则译码:02,03,12,13→02,10,14,29,24→14,21,22,31,32→21,33,34,43,44→33,00,01,40,41→40pe=05.13(1)最佳分布q(x0)=q(x1)=1/2(2)102ln)25(2ln2ln)2(2),(0qE5.14(1))1(4)ln),1(1021200ppqqqqqE-)1(21ln),1(max)1(05.00ppqEEq(2)21102002ln,1qpqqqqE)(pE121ln)1(065.15(1)])1(3)1(3[22ppppppe=(2)2232)1(3)1(2)1(533ppppppppppe(3)计算繁5.17根据最大似然译码准则,译码如下:y0→x1;y1→x1,x3;y2→x1,x3;y3→x3;y4→x1,x2;y5→x0,x1,x2,x3;y6→x0,x1,x2,x3;y7→x0;y8→x1,x2;y9→x0,x1,x2,x3;y10→x0,x1,x2,x3;y11→x0,x3;y12→x2;y13→x0,x2;y14→x0,x2;y15→x0。这样可以得到最小错误概率。5.18(1))/(322.1符号BitC(2)编码x0:00,x1:11,x2:22,x3:33,x4:44根据最大似然函数译码准则译码:00,01,10→00;11,12,21→11;22,23,32→22;33,34,43→33;40,44,04→44pe=(1/5)[(1/4)+(1/4)+(1/4)+(1/4)+(1/4)]=1/4第6章率失真编码6.5定义域:0≤D≤0.5值域:0≤R(D)≤H(X)=1.5(Bit/符号)6.6定义域:0≤D≤1值域:0≤R(D)≤H(X)=2(Bit/符号)6.7(1)选择信道105.05.001P,可使平均失真min1DD(2)选择信道为0.50.55.05.00.50.5P,可使平均失真34DDmax(说明:题有误,应为34D1Dmaxmin,)6.12DDDDijpDDDHDR11)]([21021012log)(26.132220)()()(DDHHDR反向信道DDDDyx11)]([6.15(1)R(D)=log2-H2[1-(D/β)](2)当α≥1/2,β=1时,Dmin=0,Dmax=min{/2,,/2}=1/272/102/10)(2log)(2DDDHDR6.2110102log)1()(DDDDRDDDD1001P第7章纠错编码代数基础7.1不可构成加群和乘群,无逆元7.2可构成交换加群G={0000,0001,0010,0011,0100,0101,…,1110,1111}G′={0000,0001,0010,0011}0100+G′={0100,0101,0110,0111}1000+G′={1000,1001,1010,1011}1100+G′={1100,1101,1110,1111}7.3无子群7.4构成乘群,不构成加群7.5构成加群,不构成乘群7.630个本原元,本原多项式:x5+x2+1,x5+x4+x3+x2+1,x5+x4+x2+x+1,x5+x3+x2+x+1,x5+x3+1,x5+x4+x3+x+17.7(x+1)(x5+x2+1)(x5+x4+x3+x2+1)(x5+x4+x2+x+1)(x5+x3+x2+x+1)(x5+x