Decision and risk analysis for process evolution

Abstract

Evolution is one of the integral parts in the lifecycle of a process. Similar to software, a process often has to be modified in order to fulfill changed process requirements, or to correct problems that occurred during its enactment. Although currently changes are often implemented ad-hoc without any consideration of side-effects and risks, it is obvious that such side-effects and especially the risk inherent to a change have to be analyzed before actually changing a process. Additionally, it is often impossible to stop, change and restart a process. In these cases, techniques are required which are capable of analyzing and changing currently enacted processes. The article describes the approach for risk and decision analysis elaborated in the ESPRIT project Process Instance Evolution (PIE). The main concept of the approach used in PIE consists of determining the probability distribution of important process parameters and calculating risk measures for decision making. The article describes the overall approach and explains the algorithm used for determining probability distributions.