5.12 The Hosoya Matrix

The Hosoya matrix was introduced by Randić [316]. Denoted by Z, the Hosoya matrix, is derived here in a manner similar to the edge-Wiener matrix (see section 5.3). It is a sparse symmetric square V x V matrix whose elements for a tree are defined as:

[Z]ij=
  Zi Zj             if vertices i and j are adjacent
  0                 otherwise                                       (109)      

where Zi and Zj are the Hosoya Z-indices [25] of the two subgraphs (fragments) after an edge i-j is removed from the acyclic graph. The Z-indices are tabulated for C1 to C9 fragments [93] and programs for computing this and many other molecular descriptors are available [ e.g ., 317].

Below we give the Hosoya matrix for the branched tree T2.

Z(T2)=
0
19
0
0
0
0
0
0
19
0
24
0
0
0
0
19
0
24
0
21
0
0
18
0
0
0
21
0
20
0
0
0
0
0
0
20
0
17
0
0
0
0
0
0
17
0
0
0
0
0
18
0
0
0
0
0
0
19
0
0
0
0
0
0

The Hosoya matrix may be made dense if the elements [Z]ij are computed not only for deleted edges but also for deleted edges along any path in a tree [316] . Then, the dense Hosoya matrix dZ is defined as:

[dZ]ij=
  Zi Zj             if ij
  0                 otherwise                                       (110)      

The dense Hosoya matrix for the branched tree T2 is exemplified below.

dZ(T2)=
0
19
16
12
8
4
10
11
19
0
24
18
12
6
15
19
16
24
0
21
14
7
18
16
12
18
21
0
20
10
12
12
8
12
14
20
0
17
8
4
4
6
7
10
17
0
4
4
10
15
18
12
8
4
0
10
11
19
16
12
4
4
10
0

The Hosoya matrices are used to produce a variety of molecular descriptors, especially since the Z-index and the Hosoya matrix have been extended to polycyclic systems and edge-weighted graphs [ e.g ., 318-320] .

<< . . . >>