Generating event logs through the simulation of declare models

In the process mining field, several techniques have been developed during the last years, for the discovery of declarative process models from event logs. This type of models describes processes on the basis of temporal constraints. Every behavior that does not violate such constraints is allowed, and such characteristic has proven to be suitable for representing highly flexible processes. One way to test a process discovery technique is to generate an event log by simulating a process model, and then verify that the process discovered from such a log matches the original one. For this reason, a tool for generating event logs starting from declarative process models becomes vital for the evaluation of declarative process discovery techniques. In this paper, we present an approach for the automated generation of event logs, starting from process models that are based on Declare, one of the most used declarative modeling languages in the process mining literature. Our framework bases upon the translation of Declare constraints into regular expressions and on the utilization of Finite State Automata for the simulation. An evaluation of the implemented tool is presented, showing its effectiveness in both the generation of new logs and the replication of the behavior of existing ones. The presented evaluation also shows the capability of the tool of generating very large logs in a reasonably small amount of time, and its integration with state-of-the-art Declare modeling and discovery tools.