Risk-based proactive process adaptation

Abstract

Proactive process adaptation facilitates preventing or mitigating upcoming problems during process execution, such as process delays. Key for proactive process adaptation is that adaptation decisions are based on accurate predictions of problems. Previous research focused on improving aggregate accuracy, such as precision or recall. However, aggregate accuracy provides little information about the error of an individual prediction. In contrast, so called reliability estimates provide such additional information. Previous work has shown that considering reliability estimates can improve decision making during proactive process adaptation and can lead to cost savings. So far, only constant cost functions have been considered. In practice, however, costs may differ depending on the magnitude of the problem; e.g., a longer process delay may result in higher penalties. To capture different cost functions, we exploit numeric predictions computed from ensembles of regression models. We combine reliability estimates and predicted costs to quantify the risk of a problem, i.e., its probability and its severity. Proactive adaptations are triggered if risks are above a pre-defined threshold. A comparative evaluation indicates that cost savings of upto 31%, with 14.8% savings on average, may be achieved by the risk-based approach.