Requires two groups (marked with a group column) of univariate or multivariate data, the rows in each group arranged in sequence. Sequence slotting will merge the two sequences into one, in a way that minimizes the total path length from point to point, according to the selected distance measure. A typical application is the correlation of geophysical or geochemical data from two wells or sections. The algorithm is based on Clark (1985).
An optional third group with up to three columns specifies constraints on the sequence. Each row in this group specifies one constraint using three numbers: a constraint type code followed by up to two integer parameters A and B, referring to row numbers within each original sequence.
|
Code |
A |
B |
Function |
|
1 |
s |
t |
As and Bt are near-neighbors |
|
2 |
s |
t |
As and Bt are adjacent |
|
3 |
s |
t |
As precedes Bt |
|
4 |
s |
t |
As follows Bt |
|
5 |
s |
t |
Subsequence As , Bt , As+1 |
|
6 |
s |
t |
Subsequence Bt , As , Bt+1 |
|
7 |
|
|
Sequence starts with A1 |
|
8 |
|
|
Sequence starts with B1 |
|
9 |
|
|
Sequence ends with Am |
|
10 |
|
|
Sequence ends with Bn |
|
11 |
s |
|
No B between As and As+1 |
|
12 |
|
t |
No A between Bt and Bt+1 |
Reference
Clark, R.M. 1985. A FORTRAN program for constrained sequence-slotting based on minimum combined path length. Computers & Geosciences 11: 605-617.