Modeling of an Edge Computing Cluster for Optimized Distribution of Tasks in Production Environments

Due to the digitalization of production, a vast amount of data is being generated that, through analysis and processing, offers new opportunities for implementing efficiency improvements. To process the data in real time, existing computers at the edge of production networks can be interconnected to form an edge computing cluster to distribute additional tasks among these devices. On the shop floor, satisfying latency requirements is necessary. A distribution model to optimally distribute tasks to existing computing resources is presented. The distribution model relies on a model of the network formed by the distributed edge computing nodes and their connections. Both computing nodes and their connections are characterized with various attributes. A metaheuristic approach is used to solve the optimization problem of the best distribution of tasks regarding latency and reliability.