Service Process Improvement Based on Business Process Management

Abstract

The service vendors desire to improve their service processes to retain consumers and increase profits. Most existing methods require a mass of domain knowledge to improve their service processes, which are time-consuming and error-prone. This paper proposes a method named Diff-BPI to automatically improve a service process, reducing execution cost while guaranteeing consumers’ waiting/process time. Given the existing two versions of a service process, Diff-BPI first detects their differences and all possible candidate improved service processes are constructed by connecting their same parts and combinations of different parts. After that, Diff-BPI calculates three metrics for each candidate: “quality of improvement (QoI)”, “longest execution time (LET)” and “stability of time (SoT)”, and filters the candidates with LETs higher than the given time limit and QoIs/SoTs less than the given one. Finally, Diff-BPI picks the best candidate with the lowest cost less than the costs of two existing versions. A case study shows that Diff-BPI can construct an improved service process with a lower cost than the existing two versions. The efficiency evaluation reveals that Diff-BPI can save more than 20% of the running time for improving a service process using the filter strategy when the number of differences between the two versions is more than 3.