Using monotonicity to find optimal process configurations faster

Abstract

Configurable process models can be used to encode a multitude of (different) process models. After configuration, they can be used to support the execution of a particular process. A configurable process model represents a space of instantiations (configured process variants). Such an instantiation space can be used by an organisation to select the best instantiation(s) according to some Key Performance Indicators (KPIs), e.g., cost, throughput time, etc. Computing KPIs for all the instantiations in the space is time consuming, as it might require the analysis (e.g., simulation) of thousands (or more) of instantiations. Therefore, we would like to exploit structural characteristics to reduce the amount of instantiations which need to be analysed. This reduction only removes those instantiations which do not need to be considered by an organisation. This yields the same result (a collection of best configurations), but in a faster way.