Using process-mining for understating the emergence of self-organizing manufacturing systems.

Self-organizing systems, a class of distributed systems, aim to maintain the purpose and intentions of the system regardless internal and external perturbations. These systems are composed of reconfigurable architectures and intelligent decisional entities that allow the achievement of both performance and reactivity needs. Beside other needed characteristics, such as modularity or customizability, the functioning of self-organizing systems is reliant on the degree of diagnosability during the system execution. An adequate diagnosis of the system dynamics allows the understanding the information contained and provides valuable input for the decision-making process. Process mining is a tool that permits identifying trends and patterns from event logs. This paper focuses on the use of process-mining for the diagnosis of a self-organizing manufacturing system. The approach is tested considering two self-organization rules based on the machine selection within a manufacturing environment. The approach was experimentally tested on a simulation model of a flexible manufacturing system. This exploratory research suggests that process-mining is a promising approach for the diagnosis of the behaviour of self-organizing systems.