• Italian
  • English
| Print |
Workflow systems are the core assets of the business processes of enterprises and companies. The major limitation of current workflow management systems is: low flexibility to changes and low tolerance to errors and unexpected situations. The purpose of this work packegae is to develop an integrated system for the complete management of the documentation of an enterprise of any kind and in view of the fast changing environment that the management of processes in organizations, it indicates the need for systems that workflow management are able to adapt to working dynamics according to which the initial plan could be modified, for example, to deal with the temporary unavailability of certain resources or to the change in some temporal constraints.

Planned Activities
One of the most promising models for managing decentralized workflow systems is constituted by intelligent and autonomous agent systems. These are platforms with the capacity of reasoning, based on the paradigm BDI (Belief-Desire-Intentions) that provides special functions particularly interesting for the construction of complex applications and self-organizing.
In the specific case the need to make the WFM scalable and adaptable to the particular document to be processed can be satisfied by a solution based on a combination of an expert system and a society of autonomous intelligent components. The architecture is represented in figure:
 All the sub-systems are developed through agents, with a technology-based Jason / Moise + and the FIPA standards.

In particular, the role of Expert sub-system is to classify and encode anomalied and special situations that may occur and then to develop the appropriate procedures for the proper management of the particular document. A key role will be played by the Monitor sub-system that has the task of observing the performance of the process and the work environment and to evaluate when abnormal situations occur. Record this information may be useful to the expert system in order to make predictions about future anomalies.

The implementation of the entire sub-system includes 1) a Interpreter that parses BPMN and is able to generate executable plans,  and 2) a reasoning Engine that uses BDI (belief, desire, intention) to transform plans into actions.
The agents cooperate in accordance with dynamic organizations which provide the set of actuators that interact by exchanging messages and services compatible with the standards in order to facilitate their interaction with other existing applications.

The business objectives are the expected results from the execution of a given workflow. In our view it is the enactment system to autonomously and proactively decide how to achieve these results, on the basis of the external services that are available. It is therefore necessary to decouple the activities available to the system and the business goals that can then be injected in a dynamic way.
Neither BPMN nor BPEL include any feature to specify the business objectives (although some attempts to create a business model has been made with the Business Motivation Notation (BMM). Hence, it has been necessary to model a specification language for business goals with the following features: 1) to represent the requirements and constraints of information systems, 2) to describe the expected behavior of the system, 3) a formal syntax, 4) human-oriented (focus to 'business' users), 5) to include points of uncertainty in the specification, and finally 6) it must, at least, allow you to specify any business process that can be represented in BPMN.