Deadline constrained scheduling of scientific workflows on cloud using hybrid genetic algorithm

Workflows have simplified the execution of complex large scale scientific applications. The cloud acts as an ideal paradigm for executing them but with many open challenges that need to be addressed for an effective workflow scheduling. Several algorithms have been proposed for workflow scheduling, but most of them fail to incorporate the key features of cloud like heterogeneous resources, pay-per-usage model, and elasticity along with the Quality of service (QoS) requirements. This paper proposes a hybrid genetic algorithm which uses the PEFT generated schedule as a seed with the aim to minimize cost while keeping execution time below the given deadline. A good seed helps to accelerate the process of obtaining an optimal solution. The algorithm is simulated on WorkflowSim and is evaluated using various scientific realistic workflows of different sizes. The experimental results validate that our approach performs better than various state of the art algorithms.