Based on mathematical mechanism analyses, Sections 1.2, 2.1, 2.2 strengthen the justification for the D-Quantification
methodology as well as its novelty and explore PG-Approx-Space mathematical forms. We then systematically utilize
AND OR NOT fusion technologies to establish the semantics construction and semantics granules, which underlies D-Quan-
tification and endows PG-Approx-Space with the information fusion connotation. Furthermore, we finally conduct D-Quan-
tification studies on GrC and the RS-Model. Therefore, this paper involves four concrete contents with respect to PG-Approx-
Space: (1) the basic mathematical forms, (2) the semantics construction and semantics granules, (3) the multiple granules
and their features and relationships, (4) the model expansion and model construction. Furthermore, this paper exploits a dis-
tinctive D-Quantification characteristic and has the following relevant contributions. (1) The novel mathematical forms of
PG-Approx-Space are well organized and store the double-quantitative information. (2) The semantics construction and
semantics granules systematically utilize the normal AND OR NOT technologies to complete quantitative information extrac-
tion and fusion, and the relevant system construction underlies the D-Quantification exploration. (3) GrC effectively exhibits
the rich granules and their relevant semantics, calculations and hierarchy of D-Quantification. (4) The RS-Model is system-
atically studied in PG-Approx-Space, including model expansion and model construction. Therefore, based on the quantita-
tive information architecture, this paper systematically conducts and investigates D-Quantification and particularly
establishes a fundamental and general exploration framework.
The rest of this paper is organized as follows. Regarding PG-Approx-Space, Section 2 constructs the space and plane forms
to organize and store quantitative information. Section 3 explores the semantics construction and semantics granules to ex-
tract and fuse the quantitative information. Granulation is then mainly conducted using the semantics and the microscopic
and macroscopic descriptions, and multiple fundamental granules are proposed and studied. Section 4.1 studies basic
semantics (BS) and BS-Granules; Sections 4.2 and 4.3 refer to B-Granules and C-Granules, respectively; Section 5.1 explores
M-Regions, BM-Regions, BMC-Granules; furthermore, Section 5.2 presents the granular hierarchical structure. For the RS-
Model in PG-Approx-Space, Section 5.1 proposes the model regions and model definition by developing the traditional no-
tions; Section 6.1 studies model expansion, constructs some models and further develops their structures; Section 6.2 pro-
vides a concrete model. Finally, Section 7 concludes this paper.
2. Basic mathematical forms of PG-Approx-Space
By analyzing some mathematical mechanisms on precision and grade, this section constructs the basic mathematical
forms of PG-Approx-Space, including the space and plane forms. Both forms exhibit the quantitative information structures,
rigorously organize and effectively store the quantitative information, and basically describe and construct PG-Approx-
Space. That is, they provide the mathematical basis and concrete technology for the later in-depth D-Quantification explo-
rations. These studies, in fact, concern the basic granular discussion of PG-Approx-Space because the equivalence classes are
atom-granules and their structures are ingeniously constructed herein.
2.1. Space form
Precision, internal grade and external grade can be utilized to measure the relationship between ½x
R
and A. Furthermore,
if A is treated as the set parameter while ½x
R
changes in knowledge, then the three measures can act as variables. Suppose R
denotes the real number set.
8
½x
R
2 U=R, suppose X ¼ gð½x
R
; AÞ; Y ¼ gð½x
R
; AÞ; Z ¼ pð½x
R
; AÞ. Thus, the three measures are
transposed to three variables X; Y; Z.
Definition 2.1. The three-dimensional coordinate system formed by X; Y and Z is called the measure space on precision,
internal grade and external grade and is simply noted as Measure Space.
In Definition 2.1, by being viewed as three coordinate variables, the three measures are used to establish a three-dimen-
sional coordinate system (i.e., Measure Space). In fact, the space construction uses the Cartesian product of the relative and
absolute measures, and this idea becomes scientific because D-Quantification corresponds to multiple views.
Suppose f
1
: U=R ! R
3
; f
1
ð½x
R
Þ¼ðX; Y; ZÞ. f
1
moves each equivalence class into three-dimensional Measure Space. In
other words, equivalence classes construct the three-dimensional organization using the three measures. From the three
measures perspective, Approx-Space exhibits a three-dimensional form. Moreover, PG-Approx-Space, as Approx-Space with
the three measures, is also described. We next investigate the relationship between precision and grade and further exhibit
the exact space form of PG-Approx-Space.
Proposition 2.2.
(1) Z ¼
Y
XþY
;
@Z
@X
¼
Y
ðXþYÞ
2
;
@Z
@Y
¼
X
ðXþYÞ
2
; dZ ¼
Y
ðXþYÞ
2
dX þ
X
ðXþYÞ
2
dY.
(2) X ¼ Y
1Z
Z
;
@X
@Y
¼
1Z
Z
;
@X
@Z
¼
Y
Z
2
; dX ¼
1Z
Z
dY
Y
Z
2
dZ.
(3) Y ¼ X
Z
1Z
;
@Y
@X
¼
Z
1Z
;
@Y
@Z
¼
X
ð1ZÞ
2
; dY ¼
Z
1Z
dX þ
X
ð1ZÞ
2
dZ.
X. Zhang, D. Miao / Information Sciences 268 (2014) 147–168
151