System Entity Structure (SES)  

The System Entity Structure (SES) is a high level ontology framework targeted to modeling, simulation, systems design and data engineering. Its expressive power, both in strength and limitation, derive from that domain of discourse. An SES is a formal structure governed by a small number of axioms that provide clarity and rigor to its models. The structure supports hierarchical and modular compositions allowing large complex structures to be built in stepwise fashion from smaller, simpler ones. Tools have been developed to transform SESs back and forth to XML allowing many operations to be specified in either SES directly or in its XML guise. The axioms and functionally based semantics of the SES promote pragmatic design and are easily understandable by data modelers.

Ontology is a knowledge representation concerned with describing things and their relationships. An ontology contains classes (elements), attributes of the classes, and relationships between classes with which to represent or model knowledge of a certain domain. The System Entity Structure (SES) is a formal ontology framework, axiomatically defined, to represent the elements of a system (or world) and their relationships in hierarchical manner. It provides a model to describe knowledge of a domain in a structural way. Since it is originated from the representation of simulation model structure, SES is easily accommodated in modeling and simulation for automation. While SES represents complex data in a rigorous way, it has flexibility and efficiency to change the structure according to a variety of choices. Figure 1 shows the basic representation elements of the SES.


Figure 1.   Basic SES representation


SES consists of entities, (multi-)aspects, specialization, and variables.

  • Entities represent things that have existence in a certain domain. Entities can have variables which can be assigned a value within given range and types
  • Aspects represent ways of taking things apart into more detailed ones and labeled decomposition relation between the parent and the children.
  • Multi-aspects are aspects for which the components are all of the same kind.
  • Specialization categorizes things in specific forms that it can assume. It is a labeled relation that expresses alternative choices that a system entity can take on.
  • Entities can have variables, which can be assigned a value within given range.

For example, a book can be represented in SES structure in Figure 2.

Figure 2.   Representation of a book in SES


A book consists of front cover, back cover, and pages, which show the physical decomposition relation between book and covers. The front cover of a book can be made of either cardboard or paper. The cardboard is also manufactured in red or blue. Pages contain multiple entities of the same characteristics. Pages have a variable of numOfPage whose values take on integer values.

Decomposition of event sets provides an example of the use of SES in time segment XML descriptions.

The SES operations causing structural change to extract specific information are: pruning, restructuring, and transforming. Pruning is an operation to cut off unnecessary structure in a SES based on the specification of a pragmatic frame. More specifically, it includes processes:

  1. to assign particular values to variables of entities,
  2. to trim the SES and get the minimal SES for end-users by picking specific elements from multiple choices. Restructuring is a mapping process within the same domain, and may result in the alternative structures. Transforming is also a mapping process, but from one domain to another domain. For example, a pruned SES is transformed into a simulatable DEVS model.

An important computational representation of the System Entity Structure uses the extended markup language (XML). With the availability of appropriate tool support, this makes development of XML Schema transparent to the modeler. Finally, SES structures are compact relative to equivalent Schema and automatically generate associated executable simulation models.


Natural Language Representation

The basic idea is as follows. The entity is considered as a collection of various message streams. It has been observed in complex systems that an entity node can act as receiver and sender simultaneously. It is logical to consider that a node may be processing more than one messages at a given instant. Consequently, developing a framework where the entity node model can operate with multiple message streams is the objective of this type of requirement specifications.

The rules that provide a binding to this type of requirement specifications are provided in Table 1. The designer can specify each node's behavior as a sender and a receiver with respect to any specific message type.

Rules
   
1
Copula "is" and "are" are treated the same.
2
Compounds x and y;
x, y, and z;
NOTE: the commas are mandatory for 3 or more constituents
x, y, z, and w;
x or y;
x, y, or z;
x, y, z, or w;
3
Determiners "a","the",??are removed from the input before processing
4
End of Sentence use "!" instead of "."
5
Sentence Order ??The entity mentioned first in the first sentence becomes the root of the SES.
??A variable must be attached to an entity before giving it a range specification (see below).
??Otherwise sentences can be in any order.
6
Forms In the following, CAPITALs indicate variables, lower case indicate mandatory key words in the order shown.
7
Specialization THING can be VARIANT1, VARIANT2, or VARIANT3 in CLASSFAMILY!
8
Aspect From VIEW perspective, THING is made of COMPONENT1, COMPONENT2, and COMPONENT3!
9
MultiAspect From multiple perspective, THINGS are made of more than one THING!
10
Attached Variables THING has VAR1, VAR2, and VAR3!
11
Range specifications of variables The range of THING's VAR1 is RANGE!

Table 1   Rules for Restricted Natural Language based Requirement Specifications