A handling algorithm for workflow time exception based on history logs

Time exception handling is a key technology to keep process instances running properly. For handling a kind of workflow time exceptions in advance, which is caused by the violation of deadline constraints in a process instance, an algorithm based on history logs is presented in the paper. Firstly, a time probability model of a process based on its history logs is built to model the indeterminacy of the execution time of tasks. Secondly, an algorithm is presented to detect potential time exceptions by analyzing the time probability model and the workload of current resources. To handle potential time exceptions, the task scheduling strategy adjustment (TSA) and the task execution time compression (TEC) are invoked to recover time loss. Finally, comparison experiments show that the algorithm can avoid or handle with potential time exceptions efficiently.