Cloud workflow scheduling with deadline and time slots constraints

In cloud computing, services are unlimited and available at any time from the perspective of users or tenants. However, the remaining service capability could not satisfy tenants´ requirements at any time from the perspective of service providers because the services are shared by multiple tasks. In this paper, we consider the workflow scheduling with both deadline and time slots constraints in cloud computing. An iterated heuristic framework is investigated, which mainly consists of three phases: the initial solution generation, improvement, reconstruction. Two constructive initial solution strategies, two improving methods and an reconstruction algorithm are proposed for the three phases, respectively. These strategies combine four heuristics, which are evaluated on a lot of test-bed instances.