Cloud workflow scheduling algorithm based on multi-objective particle swarm optimisation

Owing to the characteristics of market-oriented cloud computing, the objective function of cloud workflow scheduling algorithm should not only consider the running time, but also consider the running costs. The nature of cloud workflow scheduling is to map each task of a workflow instance to appropriate computing resources. Owing to the existence of temporal dependencies and causal dependencies between tasks, the scheduling of cloud workflow instance becomes more complex. The main contribution of this paper is to propose a cloud workflow scheduling algorithm based on multi-objective particle swarm optimisation. The algorithm takes makespan and total cost as two objectives. It provides users with a set of Pareto optimal solutions to select an optimal scheduling scheme according to their own preferences. The performance of our algorithm is compared with state-of-the-art multi-objective meta-heuristics and classical single-objective scheduling algorithm. The simulation results show that our solution delivers better convergence and optimisation capability as compared to others. Hence, it is applicable to solve multi-objective optimisation problems for scheduling workflows over cloud platform.