Report R2.14 Research on Workflow Management Seminar Leader: Arindam Mukherjee arindamm[at]iimcal.ac.in The following issues were discussed in the seminar: * Business processes and workflows * Existing methods for graphical representation of processes/workflows * Task Precedence Metagraphs (TPMGs) * Workflow analysis and verification using TPMGs * General issues regarding the use of graphical tools for representing complex processes Modern organizations are changing from function-based structures to process-based structures. In this context, efforts at 'business process re-engineering' seek to achieve better customer service through radical reconfigurations of work processes. Processes are sets of activities with a common goal. Business processes include material and information processes. A particular process can consist of a number of workflow sequences. A workflow is the automation of a business process wholly or partially, and consists of a set of partially ordered activities, which are undertaken, with the help of human or organizational resources. There are many methods for representing processes and workflows graphically. The WfMC process diagrams propounded by Workflow Management Coalition are widely adopted. The main focus here is on the precedence relationships between tasks. However, they do not show inputs and outputs at a particular stage. Basu and Blanning (2000) sought to answer this problem by introducing metagraphs. It is a directed graph where each node represents one or more information elements (items). An input set of items is supplied at the start. This specifies the business information initially available. Similarly, an output set is the set of items that are desired as output of the process. The metagraph approach is more appropriate than the matrix manipulation approach when processing of information is of primary interest. However, they are less suitable for the portrayal of task precedence and flow of control. Moreover, no clear distinction is drawn between OR joints & AND joins in metagraphs. Task Precedence Meta Graphs (TPMG) aim to overcome these problems. The following issues were addressed in the development of TPMGs. * Can the model have capabilities to easily verify the correctness of a process? * Can the model facilitate informational analysis, functional analysis and organisational analysis? * Can the model have the capabilities to represent temporal behaviour of the process so as to enable temporal analysis of the process such as consistency, prediction and scheduling? The architects of TPMGs sought to achieve the above objectives by the use of 'edges' and 'nodes' as the primary elements in its construction. Edges are of two types. The first is a 'task' that denotes activities to be done at a particular stage and is shown using bold arrows. The second is called a 'propagation edge' and is shown as a lightly drawn arrow illustrating the link of an item from the outgoing end of a task to the incoming end of another task. There are also two types of nodes in TPMGs. An 'init node' has a single outgoing edge corresponding to a task and is shown as a bold oval. On the other hand a 'prop node' can have multiple outgoing edges, all of which are propagation edges and are shown as lightly drawn ovals. Unlike simple metagraphs, 'and' and 'or' nodes are differentiated through various symbols in TPMGs. Various graph search algorithms could be developed for efficient use of TPMGs in workflow analysis. The seminar leader explained the mechanics of one such algorithm developed for the purpose of ascertaining whether input set 'A' was sufficient to produce an output set 'B' and whether an item 'a' was essential for producing an item 'b'. The algorithm can be extended to resolve queries related to tasks and resources and to handle temporal constraints. The verification process adopted to identify errors in process models in their design stages was also discussed. Errors can be semantic or syntactic. Semantic errors can be avoided with a clear understanding of the business processes that are being represented in the TPMG. Syntactic errors can result from the incorrect use of workflow modelling constructs. Syntactic errors such as 'deadlock', 'lack of synchronization', 'non-terminating cycles', and 'dangling nodes' were discussed in this context. The algorithm cited earlier could be enhanced to verify the syntactic correctness of a TPMG. During the course of the seminar, the similarities of information based workflows and material based workflows in several contexts were discussed. Various questions related to the use of graphical devices for representing complex, real-world processes were also raised. The uses of these graphs were identified to be two-fold. They aid communication and facilitate analysis of a business process. However, it is not easy, if at all possible to represent all the complexities of a business process graphically. At the core of a workflow analysis is the urge to standardise activities for ensuring efficiency. However, many activities are too complex for standardisation. Processes such as research and development seek to achieve innovation and their representations as routine processes will be inaccurate. Questions were also raised about the decomposability of managerial processes. It was felt that the assumption that they could be decomposed into clearly defined activities should be re-examined. Process reengineering studies typically analyse how a given process can be improved to obtain maximum efficiency. They however, do not provide answers to more strategic questions such as whether an organization should be in a particular line of business or not. References Mukherjee, A. (2004). Workflow systems: selected issues. Unpublished manuscript. Mukherjee, A., Sen, A. K., & Bagchi, A (2004). Information analysis in workflows represented as task-precedence metagraphs. Paper presented at the Fourteenth Annual Workshop on Information Technologies and Systems (WITS 2004), Washington DC, USA. Mentzas, G., Halaris, C., & Kavadias, S. (2001). Modeling business processes with workflow systems: an evaluation of alternative approaches. International Journal of Information Management, 21(2), 123-135.
Xavier Institute of Management, Xavier Square, Bhubaneswar 751013, India Research World (ISSN 0974-2379) http://www1.ximb.ac.in/RW.nsf/pages/Home |
