Improved NSGA-II based process dynamic multi-objective algorithm
Yaze Zhang · 2023
The NSGA-II-T algorithm is proposed to solve the scheduling problem with dynamic changes in process data processing in the service Internet, modelling the problem as a dynamic multi-objective optimisation problem. In the event of regular changes in load, the solution in the next environment is predicted by the transformer model and added to the original population; in the event of changes caused by physical factors such as node downtime, the population is added randomly. The algorithm uses different means of adding populations to address the lack of diversity that can occur in static multi-objective algorithms when the environment changes. Experiments conducted on the WorkflowSim platform show that the NSGA-II-T algorithm outperforms other classical algorithms in all performance metrics when the environment changes significantly, demonstrating the effectiveness of the algorithm.