張蕊蕊,陳松林
(安徽工業(yè)大學(xué) 數(shù)理科學(xué)與工程學(xué)院,安徽 馬鞍山 243032)
數(shù)值處理奇性奇攝動(dòng)邊值問(wèn)題
張蕊蕊,陳松林*
(安徽工業(yè)大學(xué) 數(shù)理科學(xué)與工程學(xué)院,安徽 馬鞍山 243032)
采用了數(shù)值積分方法求解帶有奇性的奇攝動(dòng)邊值問(wèn)題,將原邊值問(wèn)題的一般方程近似轉(zhuǎn)換為帶有極小偏差的一階微分方程,利用梯形公式得出三對(duì)角方程組,再采用修正的方法對(duì)奇性進(jìn)行處理,得出新的三對(duì)角系統(tǒng),最后利用追趕法解出三對(duì)角方程組并驗(yàn)證該方法的一致有效性.
兩點(diǎn)邊值;數(shù)值積分;奇點(diǎn);邊界層;一致有效性
奇攝動(dòng)問(wèn)題廣泛應(yīng)用在科學(xué)與工程的各個(gè)領(lǐng)域,而近年來(lái),奇異奇攝動(dòng)問(wèn)題是應(yīng)用數(shù)學(xué)家越來(lái)越感興趣的領(lǐng)域,關(guān)于奇異奇攝動(dòng)問(wèn)題的相關(guān)研究與文獻(xiàn)資料并不多。
Kadalbajoo和Reddy介紹了一種通過(guò)偏差求解奇異攝動(dòng)邊值問(wèn)題的計(jì)算方法[1],J.Ras-hidinia等得出數(shù)值方法求解兩點(diǎn)邊值問(wèn)題[2],BSL.Soujanya和G.K.Phnaeendra在前人的基礎(chǔ)上提出了一種數(shù)值積分方法去解奇性奇攝動(dòng)兩點(diǎn)邊值問(wèn)題[3]。
本文采用數(shù)值積分方法處理帶有雙奇性的奇攝動(dòng)兩點(diǎn)邊值問(wèn)題;第二部分,分別從左邊界層與右邊界層對(duì)雙奇性奇攝動(dòng)邊值問(wèn)題的數(shù)值處理方法進(jìn)行討論;第三部分,通過(guò)舉例證明本文所采用的方法是一致有效的。
考慮具有雙奇性奇攝動(dòng)兩點(diǎn)邊值問(wèn)題
其中,小參數(shù)0<μ<<1,α,β是有限常數(shù),f(x)是充分光滑的函數(shù),記
1.1 左邊界層問(wèn)題
設(shè)k1>0,在區(qū)間(0 ,1)上,取一個(gè)小的正常數(shù)0<δ<<1,假設(shè)對(duì)y"(x)在x的δ鄰域內(nèi)按泰勒級(jí)數(shù)展開(kāi),有
將(3)代入方程(1)中,則得到帶有偏差δ的一階常微分方程去近似二階微分方程(1)
其中
為了進(jìn)行數(shù)值求解,先將區(qū)間(0 ,1)等分為N個(gè)子區(qū)間,節(jié)點(diǎn)步長(zhǎng)h=1/N,xi=ih,i=0,1,…,N.在區(qū)間上對(duì)(4)進(jìn)行兩邊積分
利用梯形公式近似估計(jì)(5)式中的積分得
線性插值逼近y'(x),可得
結(jié)合(7)與(6),我們可得出方程組
在式(8)中,存在N-1個(gè)方程和N+1個(gè)未知數(shù),因此再結(jié)合(2)可以求出雙奇性奇攝動(dòng)邊值問(wèn)題的解yi,i=0,1,…,N.
1.2 右邊界層問(wèn)題
設(shè)k1<0,不難推知,奇攝動(dòng)問(wèn)題(1)和(2)在x=1的鄰域處存在邊界層。對(duì)y"(x)在x的δ鄰域內(nèi)按泰勒級(jí)數(shù)展開(kāi),有
將(9)代入(1),則二階微分方程可用帶有偏差δ的一階常微分方程去近似,得
再利用梯形公式近似估計(jì)積分,得
再由泰勒級(jí)數(shù)展開(kāi)及線性插值逼y'() x,則有
結(jié)合(12)與(13),經(jīng)計(jì)算,我們可得出方程組
但是,在i=1時(shí)Ei沒(méi)有意義,方程(14)在x=0附近具有奇性。因此,前面所獲得的差分格式系數(shù)也具有奇性。為了解決此問(wèn)題,我們將用改進(jìn)的數(shù)值方法對(duì)奇點(diǎn)作以下處理:在奇點(diǎn)x=0處,對(duì)方程(1)運(yùn)用洛必達(dá)法則,可得方程
對(duì)y"(x)在x的δ鄰域內(nèi)按泰勒級(jí)數(shù)展開(kāi)
將(16)代入(15),可得
可解得帶有右邊界層的雙奇性奇攝動(dòng)邊值問(wèn)題的解yi,i=0,1,…,N。
2.1 考慮奇攝動(dòng)問(wèn)題的左邊界層問(wèn)題
2.2 考慮奇攝動(dòng)問(wèn)題的右邊界層問(wèn)題
從上述結(jié)果來(lái)看,所得的近似解刻畫(huà)了問(wèn)題的右邊界層行為,然而由于方程的奇性較大,使得誤差相對(duì)地也變大了。
[1]Kadalbajoo M K,Reddy Y N,Numerical solution of singularly perturbation problems via deviating arguments[J], Applied Mathematics and Computation, 1987,21:221-232.
[2]Rashidinia J,Mohammadi R,Jalilian R,The numerical solution of non-linear singular boundary value problems arising in physiology[J],Applied Mathematics and Computation,2007,185:360-367.
[3]Reddy Y N,Reddy K A.Numerical integration method for general singularly perturbed two point boundary value problems[J].Applied Mathematics and Computation,2002,133(2/3):351-373.
[4]Srinivasacharya D,B.S.L.Soujanya G,Phnaeendra K. Numerical integration method for singular-singularly perturbed two point boundary value problems[J],Procedia Engineering,2015,127:545-552.
[5]顏慶津.?dāng)?shù)值分析[M].4版.北京:北京航天航空大學(xué)出版社,2006.
[6]孫志忠,吳宏偉,袁慰平,等.計(jì)算方法與實(shí)習(xí)[M].5版.南京:東南大學(xué)出版社,2011.
[7]Mohanty R K,Urvashi A.A family of non-uniform mesh tension spline methods for singularly perturbed two point singular boundary value problems with significant the numerical first derivatives[J],Applied Mathematics and Computation,2006,172:531-544.
[8]Mohanty R K,Jha N,Evans D J,Spline in compression method for the numerical solution for singularly perturbed two point singular boundary value problems [J],International Journal of Computer Mathematics, 2004,81:615-627.
Numerical treatment of singular-singularly perturbed boundary value problems
ZHANG Rui-rui,CHEN Song-lin*
(School of Mathematics&Physics,Anhui University of Technology,Ma’anshan Anhui 243032,China)
A numerical integration method is used to solve the singular-singularly perturbation boundary value problem. General equation of the original boundary value problem is transformed to an asymptotically equivalent first order differential equation with a small deviating argument.Tridiagonal system was obtained by applying Trapezoidal formula on the first order differential equation.Using the modified method to deal with the boundary value problem of singularity.Finally,we solved the tridiagonal equations by the catch-up method and verified the validity of the proposed method.
boundary value;numerical integration;singular point;boundary layer;uniformly valid
O175.6;O175.8
A
1004-4329(2017)01-004-04
10.14096/j.cnki.cn34-1069/n/1004-4329(2017)01-004-04
2016-11-15
張蕊蕊(1990- ),女,碩士生,研究方向:奇異攝動(dòng)理論。
陳松林(1964- ),男,碩士,教授,研究方向:奇異攝動(dòng)理論。Email:slchen@ahut.edu.cn。