Algorithms for Cubic Spline Interpolation

Algorithm for nding z

i
, i = 0 . . . n
% Compute the h
i
and b
i
for i = 0 : n 1
h
i
= x
i+1
x
i
b
i
= (y
i+1
y
i
)/h
i
end
% Gaussian Elimination
u
1
= 2(h
0
+ h
1
)
v
1
= 6(b
1
b
0
)
for i = 2 : n 1
u
i
= 2(h
i1
+ h
i
) h
2
i1
/u
i1
v
i
= 6(b
i
b
i1
) h
i1
v
i1
/u
i1
end
% Back-substitution
z
n
= 0
for i = n 1 : 1 : 1
z
i
= (v
i
h
i
z
i+1
)/u
i
end
z
0
= 0
How many ops are required to compute the z
i
?
Evaluating S(x)
Remember that once you have the z
i
, you can evaluate S(x) as follows:
S
i
(x) =
z
i
6h
i
(x
i+1
x)
3
+
z
i+1
6h
i
(x x
i
)
3
+ C
i
(x x
i
) + D
i
(x
i+1
x)
with C
i
=
yi+1
hi

hi
6
z
i+1
and D
i
=
yi
hi

hi
6
z
i
.
This, however, is not the most ecient computational form. We would like to use the idea of nested
multiplication, so write:
S
i
(x) = a
i
+ b
i
(x x
i
) + c
i
(x x
i
)
2
+ d
i
(x x
i
)
3
Notice that this is just the innite Taylor expansion S
i
(x) =

n=1
1
n!
(x x
i
)
n
S
(n)
i
(x
i
) (with S
(n)
i
= 0 for
n 4 since S
i
is a cubic polynomial).
Therefore,
a
i
= S
i
(x
i
) = y
i
b
i
= S

i
(x
i
) =
h
i
6
z
i+1

h
i
3
z
i
+
y
i+1
y
i
h
i
c
i
=
1
2
S

i
(x
i
) =
z
i
2
d
i
=
1
6
S

i
(x
i
) =
z
i+1
z
i
6h
i
Algorithm for Evaluating S(x)
for i = 0 : n 1
if x x
i+1
break;
end
end
h = x
i+1
x
i
Compute a, b, c and d as above.
S = a + (x x
i
) (b + (x x
i
) (c + (x x
i
)d))
How many ops are required to for each spline function evaluation?