Auto-Optimization of Business Rules Through Business Process Mining for Improved System Performance

Abstract

Data mining techniques exist that have the ability to extract useful knowledge from the available event logs. Over the years, there has been a tremendous enhancement in the IT sector. Organizations are moving towards automation of their business processes to improve and fasten their day to day business operations that in turn can lead to enhance their businesses. It is highly important for them to see their software(s)/ERP solution(s) reliable, having high performance and with full hardware resources utilization. To successfully execute business process, there can be hundreds or thousands of business rules/constraints that must be satisfied. Hard coded written sequence inside a system may reduce system performance and can lead to high resource wastages. In the present work, we first classify the business rules execution logs and then apply pattern mining techniques to predict a possible optimized business rules execution order. This new predicted optimized sequence is used for further execution of the software to adjust business rules execution order dynamically according to the running environment to make system self-adaptive. We have experimented the approach on 89304 no of failed business instances logs and found that using the proposed model, system resources consumption can be prevented from wastage by 69.7% (worst case) and 77.62% (best case) respectively.