Workflow Graph Verification Using Graph Search Techniques

Abstract

Workflow management systems provide a flexible way of implementing business processes. Structural conflicts such as deadlock and lack of synchronization are commonly occurring errors in workflow processes. Workflows with structural conflicts may lead to error-prone and undesirable results in business processes, which may in turn affect customer satisfaction, employee productivity, and integrity of data, and may also cause legal issues. Workflow verification is meant for detecting structural conflicts in workflow processes. Workflow management systems do not have the functionality for workflow verification except through simulation which does not detect the error completely. In this paper, we present a simple workflow verification method based on the principle of depth-first search. This method is meant for verifying acyclic workflow graphs. We illustrate our method with detailed workouts using business examples. We also present a detailed theoretical analysis and empirical evaluation of the proposed method. We compare our method with the well-known graph reduction based method. We observe that our method provides significantly better results. Workflow verification is crucial as workflows with structural conflicts when deployed will cause malfunctioning of workflow management systems. Moreover, our method has worst-case time complexity of O(E2) as against O((E+N)2.N2) for the graph reduction method. We believe that our method will make the workflow verification task simpler and efficient.