Graph-based multi-criteria optimization for business processes

Abstract

A process usually includes several different criteria to determine the quality of its operation. Criteria represent characteristics such as robustness, accuracy, cost and time of the complete process and of its different elements. Since there is rarely a single dominant criterion, optimization needs to evaluate multiple criteria against each other to find the most appropriate process configuration. This paper introduces a graph-based approach for the multi-criteria optimization of business processes. Based on the introduction of multi-criteria process-to-graph transitions and use-case-driven evaluation metrics, criteria graphs are created in a discrete or joint manner. Two graph evaluation types allow addressing the demands of various use cases by following an automated, priority-based iterative analysis or by analyzing in a non-strict, more comprehensive way. Originally being designed to decide on one of multiple robust process paths, the approach proves to be highly flexible for many different application areas.