1: /* @(#)j0.c 4.1 12/25/82 */ 2: 3: /* 4: floating point Bessel's function 5: of the first and second kinds 6: of order zero 7: 8: j0(x) returns the value of J0(x) 9: for all real values of x. 10: 11: There are no error returns. 12: Calls sin, cos, sqrt. 13: 14: There is a niggling bug in J0 which 15: causes errors up to 2e-16 for x in the 16: interval [-8,8]. 17: The bug is caused by an inappropriate order 18: of summation of the series. rhm will fix it 19: someday. 20: 21: Coefficients are from Hart & Cheney. 22: #5849 (19.22D) 23: #6549 (19.25D) 24: #6949 (19.41D) 25: 26: y0(x) returns the value of Y0(x) 27: for positive real values of x. 28: For x<=0, error number EDOM is set and a 29: large negative value is returned. 30: 31: Calls sin, cos, sqrt, log, j0. 32: 33: The values of Y0 have not been checked 34: to more than ten places. 35: 36: Coefficients are from Hart & Cheney. 37: #6245 (18.78D) 38: #6549 (19.25D) 39: #6949 (19.41D) 40: */ 41: 42: #include <math.h> 43: #include <errno.h> 44: 45: int errno; 46: static double pzero, qzero; 47: static double tpi = .6366197723675813430755350535e0; 48: static double pio4 = .7853981633974483096156608458e0; 49: static double p1[] = { 50: 0.4933787251794133561816813446e21, 51: -.1179157629107610536038440800e21, 52: 0.6382059341072356562289432465e19, 53: -.1367620353088171386865416609e18, 54: 0.1434354939140344111664316553e16, 55: -.8085222034853793871199468171e13, 56: 0.2507158285536881945555156435e11, 57: -.4050412371833132706360663322e8, 58: 0.2685786856980014981415848441e5, 59: }; 60: static double q1[] = { 61: 0.4933787251794133562113278438e21, 62: 0.5428918384092285160200195092e19, 63: 0.3024635616709462698627330784e17, 64: 0.1127756739679798507056031594e15, 65: 0.3123043114941213172572469442e12, 66: 0.6699987672982239671814028660e9, 67: 0.1114636098462985378182402543e7, 68: 0.1363063652328970604442810507e4, 69: 1.0 70: }; 71: static double p2[] = { 72: 0.5393485083869438325262122897e7, 73: 0.1233238476817638145232406055e8, 74: 0.8413041456550439208464315611e7, 75: 0.2016135283049983642487182349e7, 76: 0.1539826532623911470917825993e6, 77: 0.2485271928957404011288128951e4, 78: 0.0, 79: }; 80: static double q2[] = { 81: 0.5393485083869438325560444960e7, 82: 0.1233831022786324960844856182e8, 83: 0.8426449050629797331554404810e7, 84: 0.2025066801570134013891035236e7, 85: 0.1560017276940030940592769933e6, 86: 0.2615700736920839685159081813e4, 87: 1.0, 88: }; 89: static double p3[] = { 90: -.3984617357595222463506790588e4, 91: -.1038141698748464093880530341e5, 92: -.8239066313485606568803548860e4, 93: -.2365956170779108192723612816e4, 94: -.2262630641933704113967255053e3, 95: -.4887199395841261531199129300e1, 96: 0.0, 97: }; 98: static double q3[] = { 99: 0.2550155108860942382983170882e6, 100: 0.6667454239319826986004038103e6, 101: 0.5332913634216897168722255057e6, 102: 0.1560213206679291652539287109e6, 103: 0.1570489191515395519392882766e5, 104: 0.4087714673983499223402830260e3, 105: 1.0, 106: }; 107: static double p4[] = { 108: -.2750286678629109583701933175e20, 109: 0.6587473275719554925999402049e20, 110: -.5247065581112764941297350814e19, 111: 0.1375624316399344078571335453e18, 112: -.1648605817185729473122082537e16, 113: 0.1025520859686394284509167421e14, 114: -.3436371222979040378171030138e11, 115: 0.5915213465686889654273830069e8, 116: -.4137035497933148554125235152e5, 117: }; 118: static double q4[] = { 119: 0.3726458838986165881989980e21, 120: 0.4192417043410839973904769661e19, 121: 0.2392883043499781857439356652e17, 122: 0.9162038034075185262489147968e14, 123: 0.2613065755041081249568482092e12, 124: 0.5795122640700729537480087915e9, 125: 0.1001702641288906265666651753e7, 126: 0.1282452772478993804176329391e4, 127: 1.0, 128: }; 129: 130: double 131: j0(arg) double arg;{ 132: double argsq, n, d; 133: double sin(), cos(), sqrt(); 134: int i; 135: 136: if(arg < 0.) arg = -arg; 137: if(arg > 8.){ 138: asympt(arg); 139: n = arg - pio4; 140: return(sqrt(tpi/arg)*(pzero*cos(n) - qzero*sin(n))); 141: } 142: argsq = arg*arg; 143: for(n=0,d=0,i=8;i>=0;i--){ 144: n = n*argsq + p1[i]; 145: d = d*argsq + q1[i]; 146: } 147: return(n/d); 148: } 149: 150: double 151: y0(arg) double arg;{ 152: double argsq, n, d; 153: double sin(), cos(), sqrt(), log(), j0(); 154: int i; 155: 156: errno = 0; 157: if(arg <= 0.){ 158: errno = EDOM; 159: return(-HUGE); 160: } 161: if(arg > 8.){ 162: asympt(arg); 163: n = arg - pio4; 164: return(sqrt(tpi/arg)*(pzero*sin(n) + qzero*cos(n))); 165: } 166: argsq = arg*arg; 167: for(n=0,d=0,i=8;i>=0;i--){ 168: n = n*argsq + p4[i]; 169: d = d*argsq + q4[i]; 170: } 171: return(n/d + tpi*j0(arg)*log(arg)); 172: } 173: 174: static 175: asympt(arg) double arg;{ 176: double zsq, n, d; 177: int i; 178: zsq = 64./(arg*arg); 179: for(n=0,d=0,i=6;i>=0;i--){ 180: n = n*zsq + p2[i]; 181: d = d*zsq + q2[i]; 182: } 183: pzero = n/d; 184: for(n=0,d=0,i=6;i>=0;i--){ 185: n = n*zsq + p3[i]; 186: d = d*zsq + q3[i]; 187: } 188: qzero = (8./arg)*(n/d); 189: }