Collaborative, Participative and Interactive Enterprise Modeling
Abstract:
Enterprise modeling is a daunting task to be carried out from a single perspective.
A challenge to this whole complexity is conflicting descriptions given by different actors when business processes are documented.
Often enterprise modeling takes rounds of iterations and clarification before the models are verified and validated.
In order to expedite the modeling process and validity of the models, in this paper we propose an approach called Collaborative, Participative, and Interactive Modeling (CPI Modeling).
The main objective of the CPI approach is to furnish an extended participation of actors that have valuable insight into the enterprise operations and business processes.
Achieving this goal with any modeling method and language could be quite challenging.
For CPI Modeling to succeed the modeling method should adhere to certain qualities.
Next to the CPI Modeling approach, this paper discusses an enterprise modeling method that is simple, and yet powerful to capture intricate enterprise processes and simulate them.
Introduction:
Enterprise modeling is a daunting task and usually error-prone process.
The problem is not the maturity of the modelers and analysts involved, but the complex sociotechnical nature of an enterprise.
In particular, enterprise modeling in its broader context encompasses business processes, human and organizational issues, technical aspects such as information systems and enabling IT applications.
The enterprise modeling challenge could be best seen in the definition given in [14], which is defined as “computational representation of the structure, activities, processes, information, resources, people, behavior, goals and constraints of business, government, or other enterprise”.
Basically, this definition alone suffices to persuade the reader that a traditional way of enterprise modeling will not yield too much success if noninnovative approaches and methods deployed.
A traditional approach towards enterprise modeling, especially enterprise processes modeling is to delegate the work to analysts and modelers who will be normally visiting the enterprise under study, read the existing documentation, conduct a series of interviews, after which enterprise models are developed by the modelers alone.
Only several back and forth visitation and refinements will allow for a more complete model of the enterprise to emerge.
This approach has a number of problems.
As the interaction of the analysts with the enterprise employees become more and more often, the enterprise becomes more reluctant and less interested to allocate their most needed human resources to be involved in the project, which will be seen as waste of time.
This is the reason that many enterprises will consider the modeling part as hindering the IS development project.
In turn, modelers, not having sufficient rounds of iteration, will end up with a model that is either incomplete, or there are many assumptions that are intuitively made by the molders.
As a result, the model may contain a lot of flaws.
These flaws remain quite undetected as majority of enterprise process modeling is not based on formal semantics to check the models and simulate their dynamic behavior.
A comprehensive enterprise modeling requires that the models have to capture three phenomena such as enterprise processes, enterprise business rules, and enterprise information, which should be integrated in the corresponding information system deliverable [23].
Thus, it is very important that enterprise models are accurate and complete as flawed models will result in inadequate final systems, especially that system developers may rely on the models for developing an actual system as many of these complex systems are already driven by models – model-driven system development.
Traditional approaches of enterprise modeling, where modelers and analysts are the players and the rest (process owners, mangers, stakeholders, experts) are either passive participants or even absent from the scene, are least likely to result in accurate models.
Firstly, the modern enterprise business processes are too complex to be understood, captured and documented by modelers and analysts alone.
Secondly, because the resultant models should be approved by all stakeholders and decision makers, and only then the models can be implemented.
In order to address theses challenges in enterprise modeling, innovative approaches have been discussed and introduced such as participative enterprise modeling [22, 27].
A central goal of enterprise modeling is to discover domain knowledge and document the enterprise existing business processes.
The role of participative modeling is to represent this knowledge in a coherent and comprehensive model, create shared understanding, consolidate different stakeholder views, and in order to do so an extended participation of stakeholders is crucial [25].
From the foregoing, it becomes obvious that participative enterprise modeling is rather a necessity than a choice.
However, two other things that need to be considered for a successful participation are collaboration and interaction.
We refer to this approach as collaborative, participative, and interactive modeling (CPI Modeling).
It is imperative that each of the three notions is given explicit attention, which we will do in the following section.
Design and Engineering Methodology for Organizations (DEMO):
According to the DEMO theory, social actors in organization perform two kinds of acts: production act (P-acts, for short) and coordination acts (C-acts, for short).
By engaging in P-acts, the actors bring about new results or facts, e.g., they deliver service or produce goods.
Examples of P-acts are: register a student into new course; issue a ticket for a show; make a payment.
By engaging in C-acts, the actors enter into communication, negotiation, or commitment towards each other.
Examples of C-acts are: making a request for new course; presenting an issued ticket to the customer.
The generic pattern in which the two kinds of actions (P-acts and C-acts) occur is called transaction, see Figure 1 (Above Right).
In fact, a transaction is steps of C-act -> P-act -> C-act that correspondingly result in C-fact (e.g., commitment to register a student) and P-fact (e.g., do register a student).
A transaction is carried out in three phases:
the Order phase (O-phase, for short),
the Execution phase (E-phase, for short) and
the Result phase (R-phase, for short).
These 3 phases involve two actor roles.
The actor role that initiates a transaction is called initiator.
The actor role that carries out a production act is called executor.
DEMO Extended Notations:
A business transaction is a pattern of action and interaction.
An action is the core of a business transaction and represents an activity that brings about a new result.
An interaction is a communicative act involving two actor roles to coordinate or negotiate a particular action.
Each business transaction is carried out in three distinct phases (see Figure 2a, Above Left Up):
Order phase (O), during which an actor makes a ‘request’ for a service or good towards another actor.
This phase represents a number of communicative acts or interactions.
This phase ends with a commitment (‘promise’) made by the second actor, who will deliver the requested service or good.
Execution phase (E), during which the second actor fulfills its commitment, i.e., ‘produce’ the service or good.
This phase represents a productive act.
Result phase (R), during which the second actor does ‘present’ the first actor with the service or good prepared.
This phase also represents a number of communicative acts or interactions.
This phase ends with the ‘accept’ of the service or good by the first actor.
These phases are abbreviated as O, E and R correspondingly (see Figure 2b, Above Left Down).
The figure illustrates a business transaction in detailed generic form and simple OER form.
Note that the order (O) and result (R) phases are interactions and the execution (E) phase is an action.
Conclusions:
In this paper we have discussed the CPI Modeling approach and its importance for enterprise modeling, especially enterprise business processes.
We also discussed that CPI approach can be only useful if it is supported by a suitable enterprise modeling method and language that is simple, and yet powerful.
As the DutchPlast case study revealed, an enterprise process model can be very complex.
Although we omitted almost half of the case, still the processes are complex enough to be captured without collaboration of business modelers, analysts, facilitator and participation of process owners and managers.
The CPI modeling approach not only helps with the production of enterprise models and expedites the modeling process, but it allows validation and verification of the model almost immediately.
To this end, both the participation and presence of the business owners and simulation of the model can help (in the DutchPlast case we skipped simulation due to time limitation).
Application of suitable methodology is crucial to communicate among the participants.
A method that consists of a moderate number of elements, intuitive notations, and closely resembles the natural way of working in an enterprise has significant advantage as the participants do not have modeling background.
A small set of elements reduces cognitive load of the participants not familiar with the modeling method.
The modeling method should fit into the social sitting of an enterprise, which is a social system by virtue.
The discussed case study is a starting point for more extended research.
The CPI Modeling approach opens up an opportunity for many potential directions, into which the CPI modeling approach can be developed.
First of all, we intend to create a rich simulation environment around this approach and method allowing simulation of enterprise processes with animative features including building a library of customized entities.
This will create a more realistic replica of the enterprise under study.
Another interesting research topic would be to develop scenario based guidelines that will allow practitioners to apply CPI approach without extensive involvement of a facilitator.
One more interesting observation that also leads to some research was that the participants did not fully collaborate at the beginning of the CPI Modeling session.
This means that fro a full collaboration of the participants certain procedures and measures should be developed.