Particle Swarm Optimization Approach with Parameter-Wise Hill-Climbing Heuristic for Task Allocation of Workflow Applications on the Cloud

Cloud computing provides the computing infrastructure, platform, and software application services that areoffered at low cost from remote data centers accessed overthe internet. This so called "utility computing" is changingthe future of organizations in which their internal servers are discarded in favor of applications accessible in the cloud. One of the challenges workflow applications face is the appropriate allocation of tasks due to the heterogeneous nature of the cloud resources. There are different approaches, which have been proposed in the past to address the NP-complete problem of task allocation. One such approach that successfully addressed the task allocation problem made use of Particle Swarm Optimization(PSO). This paper further improves the performance of PSO by combining PSO with a local search heuristic. In particular, PSO with a parameter-wise hill-climbing heuristic (PSO-HC) for the execution of computationally-intensive as well as I/O-intensive workflows is introduced. Experiments are conducted using Amazon\´s Elastic Compute Cloud as the experimental simulation platform looking at the scalability of CPU-intensive and I/O-intensive workflows in terms of cost and execution time.