Machine-Learning based Methods for Designing and Executing Self-Adaptive Business Processes

摘要

The present invention relates to a method for the machine-learning of an execution result of a business process and a method to autonomically reflect a learning result to the execution of a post-business process. In order to design a business process, the present invention includes: a step (a) of designing a business process workflow; a step (b) of defining a variable value and a variable point based on a binding quality by defining a plurality of quality variable points to one business process, making each of the variable points have a plurality of variable values, and autonomically selecting a variable value to optimize the quality corresponding to each of the variable points; a step (c) of designing the dynamic binding of the quality variable value by enabling adaptor-based dynamic binding, corresponding to a module dynamically selecting a variable value pre-defined in accordance with quality standards, or enabling the dynamic binding of the variable value by using the confrontation of object-oriented programming; and a step (d) of forming a business process participation service. Moreover, in order to execute the business process, the present invention includes: a step (a) of tailoring the business process based on content of a variable profile with a variable value defined by variable point; a step (b) of executing the tailored business process; and a step (c) of performing machine-learning and analysis depending on a result of the execution. The present invention is capable of enabling the autonomic adaptation of the business process as well as the intellectualization of the business process, and deriving a solution to the improvement of the business process through machine-learning.