Comparison of Modelling Methodologies

K.Kosanke
CIMOSA Association e.V
kosanke@ipa.fhg.de


Business re-engineering and enterprise integration efforts are very efficiently supported by enterprise modelling methodologies. However, with the number of methodologies available the comparison and selection of the most suited one becomes a rather diff icult task. Most modelling methodologies orient themselves on the life-cycle concept but usually cover different parts of the life-cycle itself. In addition, additional terminology and sets of modelling constructs/language for representation of the modell ing constructs are further obstacles to be overcome in the selection process.

The paper (1) presents the results of an analysis carried out for the modelling methodologies ARIS, CIMOSA, GRAI/GIM, IEM and PERA, highlighting their similarities and differences. All modelling methodologies follow the enterprise life cycle - as defined in the Generalised Enterprise Reference Architecture and Methodologies Framework (GERAM) -with emphasis on the requirements definition phase. Several methodologies carry enterprise modelling through design specification and implementation description to ope ration and model maintenance. More details are shown in Table 1 , the full analysis is contained in the original paper.

GERAM ARIS CIMOSA GRAI/GIM IEM PERA
Identification not defined not defined not defined not defined EBE (2) Identfication
Concept not defined not defined not defined not defined EBE Concept Layer
Requirement Operation Concept Requirement Definition Concept Level Analysis Requirement Definition EBE Definition Layer
Design IT System Concept Design Specification Structure Level User Oriented Design System Design EBE Specification Layer
EBE Detailed Design Layer
Implementation Implementation Implementation Description Realisation Level, Technical Oriented Design Implementation Description EBE Manifestation Layer
Operation (Operation) EBE Operation Layer
System Change Model Maintenance Model Update

Table 1 : Modelling Framework Comparison - Life Cycle (Modelling Levels)

Language expressiveness is quite different both in number of language constructs provided and in their use for enterprise modelling. The modelling languages have been compared and mapped against the European Pre-Standard ENV 12 204. A comparison of the constructs of the modelling languages is shown in Table 2.

Modelling Constructs ENV 12 204 ARIS CIMOSA GRAI/GIM IEM PERA
General Definitions not defined not defined Engineering Environment, Operation Environment Decis. System Inform. Syst. Phys. System, Bus. Domain not defined Enterprise Business Entity
Function View / Static Enterprise Activity Function Domain, Enterprise Activity (Funct.Oper.) IDEF0 Activity Activity, Function, (Action) Task Module
Function View / Dynamic Bus. Process, Event, (Sequential Relationships) Process Chain, Event, (Connectors), Cluster Process (DP, BP), Event, (Behav.Rules) not defined Funct. Chain, Funct Auton. Unit, (Connect Constructs) not defined
(Decision ) View not defined not defined not defined GRAI Grid: Decision Level/ Centre/ LinkGRAI Net: Decision./ Not Dec Activity not defined not defined
Organisation View Organisational Unit Organ. Level, Organ. Unit, Attribute, Location, Network, NetworkNode, NetworkUnit, TechResource Organ. Cell, Organ. Unit, Organ. Element not defined Object Class: Special Resource not defined
Information View Enterprise Object, Product, Order, Object View, Relation Entity, Attribute, Relation, Terminology, Table, (Cardinality, Operators) Enterprise Object, Object View, (Inf. Element), Relation, (Cardinality, Operators) Information Model, Entity, Relation Object Class: Product, Order , Relation (Operators) not defined
Resource View Capability Set, Resource part of Organisation View Capability Set, Resource / Functional Entity not defined Object Class: Resource not defined
Number of Constructs 11 17 11 9 10 1

Table 2 : Modelling Language/ Construct Comparison

In addition, the business process representation provides explicit identification of the information to be collected in the model. Both the information needed for the different modelling tasks and the results of the tasks can be explicitly identified thereby guiding the user of the methodology. The analysis identifies the compatibilities of the different modelling methodologies and their emphasis on particular parts of the enterprise modelling task. It demonstrates the value of process oriented representation of modelling methodologies. Most importantly the process model allows to identify the information used an produced during model creation. This information will lead to a consistent knowledge base of the enterprise in the course of enterprise modelling.

Additional benefits of the analysis will be obtained by taking advantage of the common representation and by converging terminology and task definitions. There is no common understanding on enterprise models and relating models from different enterprises is a rather difficult if not impossible task. It is the intent of this work to help to harmonise the results of enterprise modelling as well as the terminology used. Both are very much needed to improve enterprise integration leading to co-operations acro ss organisational boundaries required in extended or virtual, agile enterprises.


1. The original paper will be published by Chapman & Hall in the Conference Proceedings of the DIISM'96
2. EBE (Enterprise Business Entity)

www-cimosa@cnt.pl, 21.10.1996 (last update: 21.10.1996)