PRAJNA ~ Joumal of Pure and Applied Sciences Vol. 21: 30 - 34 (2013) ISSN 0975 - 2595 ERROR FUNCTIONS AND THEIR APPLICATION ‘Bhailal P. Patel, “Jyotindra C. Prajapati ‘Department of Mathematics, N.V: Patel College of Pure and Applied Sciences, Vallabh Vidvanagar-388 120 *Department of Mathematics, Marwadi cation Foundation Group of Insitutions, Rajkor-. 0.003 ABSTRACT In this paper, authors studi incomplete Gamma fun J origin of error function, its series form, its relationship with Confluent Hypengeometric funct 1s. Laplace transform of error function and its generaliza mn have been discussed. Differentiation and integration of error function and an engineering application of error function also been obtained. 2010 MSC Subject Classification: 33B15,33B99, 33520 Keywords: error fnction, complementary error fiction, Gamma finetion Introduction and Preliminaries Error function is used in measurement theory (using probability and statistics), and although its use in other branches of mathematics has nothing to do with the characterization of measurement errors, the name has stuck. The Error function is a function which occurs in probabilily, statistics and partial differential equations. Error function plays an important role in the theory of the normal random variable and probability determination, In the present paper we compute the nth derivative of the error function in terms of Hermite function by using Power series representation of the error function, which is also used to discuss the relation of error funetion with confluent Hypergeometric function and incomplete gamma function, We compute the Laplace transform of the error function and suggest a generalization of the error function, We solved one dimensional Initial boundary Value problem for one dimensional diffusionequation in terms of error function. Gamma function (Rainville[1}) is defined as To = [Pe We 'dt, Re(a)>0 wo Incomplete Gamma function (Rainville{}) of fistkindis ya,x)= [Petr tar @ Incomplete Gamma function (Rainvlle|}) of second kind defined as Te.x=feer a 8) Error function (Rainville| I) defined as 2 and complementary error function (Rainville{1)) defined as efx exp(-P yd, Xisareal(§) ere) fear r a isa) AHypergeomettic function pFq (Rainville{1)) definedas ‘Comssponding author: fppatal Vignal © where no denominator parameter bq is allowed tobe zero ora negative integer, Kummer’s Theorem (Rainville[ 1) states is neither zero nornegative integer then VF (asbix)= 6° \F, (b= a:b:-x) Laplace transform (Rainville) of £() defined LFW}= fe“renat, Rewsy> 0 ® Fourier trinsform (Lokenath Debnath[2]) of f(x) isdenotedby Fin) (), ke and defined by the integral Furent= e*read: (9) Inverse Fourier transform (Lokenath Debnath[2)) of fix) is denoted by FUP, = Sis defined by FYFUD=f0)= Ee feF& (10) Fourier sine transform (Lokenath Debnath{2) defined by Farin cee fEffainte renee Hermite polynomials Hn(x) and its recurrence relations ‘ainville[1} defineas Dime Hoy=% CDenan 0° Far 2! Hi (x)= 2nH , (x) (2)31 PRAJNAsJournal of Pure and Applied Sciences, Vol. 21: 30 - 34 (2013) H(9)= 28H, (9) ~2in-1) Hy 465) (3) Rodrigues’ formula for Hermite polynomials (Raimville[1) givenby HC) = (-D" exp (2 )D "exp Cx*) ‘The Error Function ‘The Normal distribution was first introduced by: {de Moivre in 1733 inthe context of approximating certain Binomial distributions forlarge n, His result was extended by Laplace in his book Analstical Theory of Probabilities (1812) and is now called the Theorem of de Moivre- Laplace. Laplace used the normal distribution in the analysis of errors of statistical distribution. The important ‘mumerical method of least squares was introduced by Legendre in 1805. Well-known Normal distribution tisineas 1 Ifx-w eo (=u) | (OTe mas }} ay —w0 Considering 4~Oando~ 1, thestandad normal isuiboondetinl — lel, wcrc f= geenl-de}: as Gauss considered the integral = fens hat ban 1=fqeeo{-fe | as) Aboveintegral canbe writenas F=f ewer it ay Gauss studied this integral inthe range -x tox ‘and he found that this integral is the error of normal distribution in terms of functionas fee spt) Fin hen ee as) Figure 1: Graph of error function December By considering integral (17) inthe range >to oand usingdefinition of Gamma function, we get Pie-rya ve ie. Lemerx Tscanbe stn #76 ett yar+ [ewer yd )=1 ‘his gives the definition of complementary error function, as 2 ¥ [Pexpr yak Figure 2: Graph of complementary error function Elementary Properties of error function Lerf(o) =I 2erf@)=0 Bef) 4. erfe(0)=1 S.erfelo) 6 erf(-x)=-erf(x) 7. erfe(-x) = 1+ erf (x) B.erfe—x)+erfe(x) =2 Series Form of Error Function Since the series for e”" converges absolutely and ‘uniformly forall te R, we have2013 offer 2 pseu “gle (20) 2507" Few oa In view of the convergence behavior of th series, we further have term by term differentiation of (20) fen respect ons we get nba Deel an Foe *) ‘Third differentiation gives Zeerny= @ 2 & Now, we will derive following expression for n® derivative of error function by using mathematical inductiononn, 4a ove 1-22," * 23) ae —De* ef) = wan, aee* 24) ro ‘where H, isthe Hermite polynomial From equations (21) t0 23), result (24) istrue for 1 1, 2,3. Assume that (24) holds for n-K. So by induction hypothesis and using (12) and (13), we get a greg 4s costs] a cy SS 91 yt ° eo” Pu Dit, ,09- 241, 9] yt" al Beth e)-20-DH 0] DY Bethe . ‘Therefore. by mathematical induction equation (24) holds. Using integration by parts, we get following Bhailal P. Patel et al. - Error Functions and their Application 32 (ertaasepeyebee!-n 290 08 Perewnt renee ted-e"), 250 26) @ 28) ‘The Series form of an error function (Taylor's seriesat0)is eft= Bex see tee tee L 29) here the series in right hand side is absolutely. and uniformly convergent, Hence 2 i 70 82 | «0 Relationship With Confluent Hypergeometric Fonction Fromequation (20), we get ens) ep Using Kummer’ theorem (7), we get 2x 13.5 eee alge i) @) Relationship With Incomplete Gamma Functions Incomplete Gamma functionof first kind is etd =2fte 3)33 PRAJNAsJournal of Pure and Applied Sciences, Vol. 21: 30 - 34 (2013) ‘The incomplete Gamma function of second kind Vi erfetx)= VR -erft) G4) Laplace Transform of an Error Function From definition of Laplace Transform Her oin= fe" Se dt ‘Since e“ is bounded and continuous for every 1(-92<1<2°), changing he order of integration we get ple Levee -#t [Seu By takingu+s/2 eget Mert isl= ee! fre a ‘This reduces to, - Herfhs) = beer) 65) Similarly, the Laplace Transform of erfit) is obtainedas: Lerf?ys\= fe “eh ef dudt ‘As_¢" is bounded and continuous for every 1-69 <1 <°°).by changing the order of integration, we get 5 flew fematde 2p eto yeh Heh )s Taking substitution (1 +s)u"=x' ,wehave Hence, Merfry:s AGeneralization sults a A generalization of error function (4)is December, Foon plete G7) where n= 0,1,2,..... The following results are {immediately follows rom (37), se (3) Lege seth ad Brak 1 Suse Eka Figure3: Graphof x) case ‘An Application Consider, the iital-boundary value problem for ‘the one-dimensional diffusion equation with no sources or sinks OH pene 1 38) Soe 150 68) where isa diffusivity constant, with the initial condition u(s,0)= f(s), “ewer 6») ‘This problem can be transform in the following system 5 U,=-KkU, 10, Ue =F) ‘The solutionof theubovesystemis Udo Fae” ‘The inverse Fourier transform gives the solution x)= Fe J Reese wi oat ap ‘Applying Convolution theorem, woxd= ef sesso 2) where,2013 Bhailal P. Patel et al. - Error Functions and their Application 34 L x pa he (43) aaitey ‘Therefore, solution (42) reduces to ay won= ef seed 2S Using change of variable ae a 2k Wet toexpress solution (44) inthe form. nas (45) H0= Ef see vee ot This integral is convergent for all time t> 0 and, the imegrals obtained from (44) by differentiation under the integral sign with respect to X and t are uniformly convergent in the neighborhood of the point (x.t). Hence. the solution u(s.t) and its derivatives of all orders exist for Po. We consider a special case_involving discontinuous initial condition in the form f(x) =T.H(s) where T, is constant. Inthiscase solution (44) becomes. fren wenn Speeloon, By puting "ORE (46) ‘solution (46) can be express in the form of error function © sone fel etn Ber( 55) Elrron(s)] Acknowledgement Authors are grateful to reviewer for their ‘valuable suggestion and comments for betterment of paper References [1] Rainville E.D.: Special Functions, Macmillan, New ‘York, 1960, [2] Lokenath Debnath: Intrgral Transforms and their applications, second edition, Chapman and Hall/ ‘CRC Taylorand Francis Group. 2007, [3] Srivastava H.M. and Manocha H.L.: A treatise on ‘generating functions, John Wiley and Sons, New. ‘York, Ellis Horwood, Chichester, 1984, [4] Lebedev: Special Functions and their applications, Prentice-Hall, (1965), Englewood Cliffs, NJ. [5] Mclachlan : Function of Complex Variable (1963), ‘The syndics of Cambridge Uni, Press. [6] "Erf." From MathWorld--A Wolfram Web Resource hntp:/mathworld.wolfram.com/Erfhimt