Coordinator(s): Kerry Hinge, Aditya Ghose and George Koliadis
ProcessSEER (Semantic Effect Evaluation and Reasoning) is a decision making assistant that will automatically generate an unambiguous set of requirements for the development of semantic web services.
The project seeks to develop a tool suite that can integrate the efforts of management, process analysts and application developers and assist with the coordination of diverse and sometimes conflicting motivations that influence enterprise design decisions. Currently, many process changes within an enterprise involve optimising process models. These optimisations may appear logical within the context of the modelling framework but in practice may have dire consequences with regard to productivity. ProcessSEER allows managers and analysts to change business process models and receive immediate feedback about the consequences of those changes without the need for real-life testing. This can have significant cost saving benefits with regard to process optimisation errors.
ProcessSEER also links directly with semantic web services via a semantic web service description language file (e.g. WSDL-S, OWL-S or WSMO) that acts as a template or requirements document upon which application developers can build the corresponding applications. The same semantic description file can also be imported into ProcessSEER as an activity that can be inserted into a business process model.
The tool suite consists of a Business Process Modelling Notation (BPMN) modelling tool, a Controlled Natural Language (CNL) translator, a reasoning engine, an ontology editor and an interface for communicating directly with online semantic web services. The modelling tool currently allows analysts to annotate process model elements with semantic effects. The semantic effect annotations can be accumulated across a process model with consistency checking performed at each activity within the model. The modelling tool currently requires that the analyst enter semantic effect annotations in CNL. The CNL translator facilitates annotation in natural language while the ontology editor allows the analyst to selectively classify the terminology used in the semantic effect annotations. Finally, the interface with semantic web services allows models to be directly linked with those services via their activities.
The tool suite is built using the Eclipse platform as its foundation. The image above shows the interface of the modelling tool which utilises the SOA Tools Platform (STP). At the bottom of the interface is an input field for immediate effect annotations that are associated with individual activities in the model. Directly below that is the automatically calculated cumulative effect field.
There can be many different routes by which to arrive at a selected activity in a process model. Each route constitutes a scenario label, i.e. a sequence of preceding activities performed that result in the selected activity. An effect scenario is the accumulation of all the immediate effects of each activity in a scenario label and the cumulative effect is the collection of all effect scenarios that result in the selected activity. In the case of the model above, the activity T12 is selected and the corresponding effect scenarios are displayed in the cumulative effect field at the bottom of the screen.