Knowledge Graph-Enhanced Digital Twin Framework for Optimized Job Shop Scheduling in Smart Manufacturing

The emergence of Digital Twin (DT) technology over the past decade has introduced a transformative perspective to various research areas, notably the Job Shop Scheduling Problem (JSSP) in smart manufacturing. This paper proposes a novel algorithm to address the JSSP in industrial contexts, operating under different modes to minimize makespan and calculate energy consumption. The approach is grounded in a DT framework that leverages a knowledge graph to represent the assets of the manufacturing system. Two algorithms are designed to balance the time required to complete a list of jobs with the energy consumed by the machines in the factory, all while considering the delivery deadline of the final product. The first algorithm is a genetic algorithm (GA) that minimizes the makespan required to complete the job list for each mode of operation followed by a grey wolf optimization algorithm (GWO) to verify and enhance the results of the GA. The second algorithm uses the results of the first algorithm to calculate the running time and energy consumption for each machine in each production line, considering that the production lines of the factory vary in capacity and power, then computes the total energy consumption and presents the results to the decision maker via a dashboard. The dashboard displays, for each mode of operation, the makespan, total energy consumption, and the deadline for all the jobs. To validate the efficiency and accuracy, a case study of a Smart Poultry Feed Planet (SPFP) has been conducted.