3-Lecture-Special Functions MAT - M Sc-IV
3-Lecture-Special Functions MAT - M Sc-IV
3-Lecture-Special Functions MAT - M Sc-IV
Afzal Rana
Lecture-III
NONPOLYNOMIAL COEFFICIENTS:
➢ The following example illustrates how to find a power series solution about the ordinary
point 𝑥0 = 0 of a differential equation when its coefficients are not polynomials.
➢ In this example we see an application of the multiplication of two power series.
𝑦(𝑥) = ∑ 𝑐𝑛 𝑥 𝑛
𝑛=0
𝑛=2
𝑥2 𝑥4 𝑥6
(2𝑐2 + 6𝑐3 𝑥 + 12𝑐4 𝑥 2 + 20𝑐5 𝑥 3 + ⋯ ) + (1 − + − + ⋯ ) (𝑐0 + 𝑐1 𝑥 + 𝑐2 𝑥 2 + ⋯ ) = 0
2! 4! 6!
pg. 1
Special Functions MAT 4006 M Sc-IV Prof. Dr M. Afzal Rana
1 1
2𝑐2 + 𝑐0 + (6𝑐3 + 𝑐1 )𝑥 + (12𝑐4 − 𝑐0 + 𝑐2 ) 𝑥 3 + (20𝑐5 − 𝑐1 + 𝑐3 ) 𝑥 3 + ⋯ = 0.
2 2
It follows that
1 1
2𝑐2 + 𝑐0 = 0, 6𝑐3 + 𝑐1 = 0, 12𝑐4 − 2 𝑐0 + 𝑐2 = 0, 20𝑐5 − 2 𝑐1 + 𝑐3 = 0, and so on.
This gives
1 1 1 1 1 1
𝑐2 = − 2 𝑐0 , 𝑐3 = − 6 𝑐1 , 12𝑐4 = 2 𝑐0 − 𝑐2 = 2 𝑐0 + 2 𝑐0 = 𝑐0 ⟹ 𝑐4 = 12 𝑐0 ,
1 1
20𝑐5 = 2 𝑐1 − 𝑐3 ⇒ 𝑐5 = 30 𝑐1 , ⋯. Thus, the series solution
𝑦(𝑥) = 𝑐0 + 𝑐1 𝑥 + 𝑐2 𝑥 2 + 𝑐3 𝑥 3 + 𝑐4 𝑥 4 + 𝑐5 𝑥 5 + ⋯,
yields
1 1 1 1
𝑦(𝑥) = 𝑐0 + 𝑐1 𝑥 − 𝑐0 𝑥 2 − 𝑐1 𝑥 3 + 𝑐0 𝑥 4 + 𝑐1 𝑥 5 + ⋯,
2 6 12 30
1 1 1 1
𝑦(𝑥) = 𝑐0 (1 − 𝑥 2 + 𝑥 4 − ⋯ ) + 𝑐1 (𝑥 − 𝑥 3 + 𝑥 5 − ⋯ ),
2 12 6 30
𝑦(𝑥) = 𝑐0 𝑦1 (𝑥) + 𝑐1 𝑦2 (𝑥),
where
1 1
𝑦1 (𝑥) = 1 − 𝑥 2 + 𝑥 4 − ⋯,
2 12
1 1
𝑦2 (𝑥) = 𝑥 − 𝑥 3 + 𝑥 5 − ⋯.
6 30
Because the differential equation has no finite singular points, both power series converge for
|𝑥| < ∞.
➢ These equations are similar only in that they are both examples of simple linear second-
order DEs with variable coefficients.
➢ Since 𝑥 = 0 is an ordinary point of Eq. (2.2.1), therefore, there is no problem, as
discussed in previous section, in finding two distinct power series solutions centered at
that point.
pg. 2
Special Functions MAT 4006 M Sc-IV Prof. Dr M. Afzal Rana
➢ In contrast, because 𝑥 = 0 is a singular point of the Eq. (2.2.2), finding two infinite
series—notice that we did not say power series—solutions of the equation about that
point becomes a more difficult task.
is further classified as either regular or irregular. The classification again depends on the
functions 𝑃 and 𝑄 in the standard form
𝑦 ′′ + 𝑃(𝑥)𝑦 ′ + 𝑄(𝑥)𝑦 = 0, (2.2.4)
where
𝑎1 (𝑥) 𝑎0 (𝑥)
𝑃(𝑥) = , 𝑄(𝑥) = .
𝑎2 (𝑥) 𝑎2 (𝑥)
If 𝑥 = 𝑥0 is a regular singular point of the differential equation (2.2.3), then there exists at least
one solution of the form
𝑦 = (𝑥 − 𝑥0 )𝑟 ∑∞ 𝑛 ∞
𝑛=0 𝑐𝑛 (𝑥 − 𝑥0 ) = ∑𝑛=0 𝑐𝑛 (𝑥 − 𝑥0 )
𝑛+𝑟
, (2.2.5)
where the number 𝑟 is a constant to be determined. The series will converge at least on some
interval 0 < 𝑥 − 𝑥0 < 𝑅.
➢ Notice the words at least in the first sentence of Theorem 2.2.1. This means that in
contrast to Theorem 2.1.1, Theorem 2.2.1 gives us no assurance that two series solutions
of the type indicated in (2.2.5) can be found.
➢ The method of Frobenius, finding series solutions about a regular singular point 𝑥0 , is
similar to the method of undetermined series coefficients of the preceding section in that
pg. 3
Special Functions MAT 4006 M Sc-IV Prof. Dr M. Afzal Rana
we substitute 𝑦 = ∑∞𝑛=0 𝑐𝑛 (𝑥 − 𝑥0 )
𝑛+𝑟
into the given differential equation and determine
the unknown coefficients 𝑐𝑛 by a recurrence relation.
➢ However, we have an additional task in this procedure: Before determining the
coefficients, we must find the unknown exponent 𝑟.
As we did in the discussion of solutions about ordinary points, we shall always assume,
for the sake of simplicity in solving differential equations that the regular singular point is 𝑥0 .
Now
∞ ∞
′ 𝑛+𝑟−1
𝑦 = ∑ 𝑐𝑛 (𝑛 + 𝑟)𝑥 , and 𝑦′′ = ∑ 𝑐𝑛 (𝑛 + 𝑟)(𝑛 + 𝑟 − 1)𝑥 𝑛+𝑟−2 .
𝑛=0 𝑛=0
or
∞ ∞ ∞
𝑛+𝑟−1 𝑛+𝑟−1
3 ∑ 𝑐𝑛 (𝑛 + 𝑟)(𝑛 + 𝑟 − 1)𝑥 + ∑ 𝑐𝑛 (𝑛 + 𝑟)𝑥 − ∑ 𝑐𝑛 𝑥𝑛+𝑟 = 0,
𝑛=0 𝑛=0 𝑛=0
or
∞ ∞
𝑛+𝑟−1
∑ 𝑐𝑛 (𝑛 + 𝑟)(3𝑛 + 3𝑟 − 2)𝑥 − ∑ 𝑐𝑛 𝑥 𝑛+𝑟 = 0,
𝑛=0 𝑛=0
or
∞ ∞
𝑟 𝑛−1
𝑥 [∑ 𝑐𝑛 (𝑛 + 𝑟)(3𝑛 + 3𝑟 − 2)𝑥 − ∑ 𝑐𝑛 𝑥 𝑛 ] = 0,
𝑛=0 𝑛=0
pg. 4
Special Functions MAT 4006 M Sc-IV Prof. Dr M. Afzal Rana
or
∞ ∞
𝑟 −1 𝑛−1
𝑥 𝑐0 𝑟(3𝑟 − 2)𝑥 + ∑ 𝑐𝑛 (𝑛 + 𝑟)(3𝑛 + 3𝑟 − 2)𝑥 − ∑ 𝑐𝑛 𝑥 𝑛 = 0,
⏟
𝑛=1 ⏟
𝑛=0
[ 𝑘=𝑛−1 𝑘=𝑛 ]
∞ ∞
𝑟 −1
𝑥 [𝑐0 𝑟(3𝑟 − 2)𝑥 + ∑ 𝑐𝑘+1 (𝑘 + 𝑟 + 1)(3𝑘 + 3𝑟 + 1)𝑥 − ∑ 𝑐𝑘 𝑥 𝑘 ] = 0,
𝑘
𝑘=0 𝑘=0
∞
𝑟 −1
𝑥 [𝑐0 𝑟(3𝑟 − 2)𝑥 + ∑{𝑐𝑘+1 (𝑘 + 𝑟 + 1)(3𝑘 + 3𝑟 + 1) − 𝑐𝑘 }𝑥 𝑘 ] = 0,
𝑘=0
Equation (2.2.8) is called the indicial equation of the problem, and its roots are called the
indicial roots, or exponents, of the singularity 𝑥 = 0. From Eq. (2.2.8), we have the following
indicial roots
2
𝑟 = 3 , and 𝑟 = 0. Thus, two different recurrence relations are:
2
When 𝑟 = 3, When 𝑟 = 0
𝑘 𝑐 𝑐
𝑘
𝑐𝑘+1 = (3𝑘+5)(𝑘+1) , 𝑘 = 0,1,2, ⋯. 𝑐𝑘+1 = (𝑘+1)(3𝑘+1) , 𝑘 = 0,1,2, ⋯.
0 𝑐 0 𝑐 0𝑐
𝑐1 = 5∙1 = 1!5 𝑐1 = 1∙1
1 𝑐 0 𝑐 1𝑐 0𝑐
𝑐2 = 8∙2 = 2!5∙8 𝑐2 = 2∙4 = 2!1∙4
2 𝑐 0 𝑐 2𝑐 0 𝑐
𝑐3 = 11∙3 = 3!5∙8∙11 𝑐3 = 3∙7 = 3!1∙4∙7
3 𝑐 0 𝑐 𝑐
3 0 𝑐
𝑐4 = 14∙4 = 4!5∙8∙11∙14 𝑐4 = 4∙10 = 4!1∙4∙7∙10
pg. 5
Special Functions MAT 4006 M Sc-IV Prof. Dr M. Afzal Rana
⋮ ⋮
𝑐0 0 𝑐
𝑐𝑛 = 𝑛!5∙8∙11⋯(3𝑛+2) 𝑐𝑛 = 𝑛!1∙4∙7⋯(3𝑛−2)
Here we encounter something that did not happen when we obtained solutions about an ordinary
point; we have what looks to be two different sets of coefficients, but each set contains the same
multiple 𝑐0 . If we omit this term, the series solutions are
𝑛=0 𝑛=0
∞ ∞
2⁄ 2⁄ 𝑐0
𝑦1 (𝑥) = 𝑥 3 [𝑐0 + ∑ 𝑐𝑛 𝑥𝑛 ] = 𝑥 3 [𝑐0 +∑ 𝑥𝑛]
𝑛! 5 ∙ 8 ∙ 11 ⋯ (3𝑛 + 2)
𝑛=1 𝑛=1
∞
2⁄ 1
𝑦1 (𝑥) = 𝑐0 𝑥 3 [1 +∑ 𝑥𝑛]
𝑛! 5 ∙ 8 ∙ 11 ⋯ (3𝑛 + 2)
𝑛=1
∞
2⁄ 1
𝑦1 (𝑥) = 𝑥 3 [1 +∑ 𝑥 𝑛 ] (omitting 𝑐0 ).
{ 𝑛! 5 ∙ 8 ∙ 11 ⋯ (3𝑛 + 2) }
𝑛=1
2⁄ 1
𝑦1 (𝑥) = 𝑥 3 [1 + ∑∞ 𝑛
𝑛=1 𝑛!5∙8∙11⋯(3𝑛+2) 𝑥 ], (2.2.10)
1
𝑦2 (𝑥) = 𝑥 0 [1 + ∑∞ 𝑛
𝑛=1 𝑛!1∙4∙7⋯(3𝑛−2) 𝑥 ]. (2.2.11)
By the ratio test it can be demonstrated that both (2.2.10) and (2.2.11) converge for all values of
𝑥 −that is, ⌊𝑥⌋ < ∞. Also, it should be apparent from the form of these solutions that neither
series is a constant multiple of the other, and therefore 𝑦1 (𝑥) and 𝑦2 (𝑥) are linearly independent
on the entire 𝑥 −axis. Hence by the superposition principle,
pg. 6
Special Functions MAT 4006 M Sc-IV Prof. Dr M. Afzal Rana
Remarks:
➢ In general, after substituting 𝑦 = ∑∞ 𝑛=0 𝑐𝑛 𝑥
𝑛+𝑟
into the given differential equation and
simplifying, the indicial equation is a quadratic equation in 𝑟 those results from equating
the total coefficient of the lowest power of 𝑥 to zero.
➢ We solve for the two values of 𝑟 and substitute these values into a recurrence relation
such as (2.2.9). Theorem 2.2.1 guarantees that at least one solution of the assumed series
form can be found.
pg. 7