Corrective Evolution of Adaptable Process Models

The aim of corrective evolution is to update a process model by plugging in successful process instance adaptations, while ensuring that the resulting model satisfies the goal of the original model. Corrective evolution is necessary if adapting at runtime is costly, or if adaptations fail and should be avoided. Considering that a trace is a recording of executed activities, we identify three possibilities for correcting process models, depending on the traces on which the instance adaptation should be plugged in. A correction is strict if the adaptation should be plugged in on a precise trace, relaxed if on all traces, and relaxed with conditions if on a subset of all traces. Automated techniques for corrective evolution are necessary since changing models manually is difficult and there is a need to verify that the evolved model satisfies the goal. We provide automated solutions for two cases: when corrections are strict, and when they are either strict or relaxed. We evaluate the tradeoffs between strict and relaxed corrections using a real log.