An Improved Scheduling Algorithm for Dynamic Batch Processing in Workflows

Aiming at the shortcomings in existing scheduling methods for dynamic batch processing in workflows, a new scheduling optimization model considering varied factors such as the resource´s competence and execution difficulty of activity instances is established. Consequently, a multi-objective optimization algorithm based on the theory of particle swarm optimization, MOPSO-TC, is proposed. The effectiveness of the MOPSO-TC algorithm is evaluated by comparing its results to the multi-objective particle swarm optimization with the sigma method (SMOPSO) and the time variant multi-objective particle swarm optimization (TV-MOPSO). The experimental results indicates that the MOPSO-TC algorithm reports better quality solutions on different problem instances.