A New Shuffled Genetic-based Task Scheduling Algorithm in Heterogeneous Distributed Systems
Distributed systems such as Grid- and Cloud Computing provide web services to their users worldwide. One of the most important concerns that service providers face is total cost of ownership (TCO). A large portion of TCO is related to power consumption due to inefficient resource management. Task scheduling module as a key component can have a great impact on user response time and underlying resource utilization. Such heterogeneous distributed systems have used different processors with different speeds and architectures. Also, the user program, which is usually represented as a directed acyclic graph (DAG), must be executed on these types of parallel processing systems. Since work scheduling in such complex systems is part of NP-hard problems, the existing heuristic approaches are no longer efficient. Therefore, the workflow is to use hybrid meta-heuristic approaches. In this paper, we have presented a meta-heuristic genetic shuffled task scheduling algorithm to minimize the total execution time and duration of user programs. In this regard, we take advantage of other heuristic methods such as Heterogeneous Fastest Termination Time (HEFT) to generate an intelligent initial population using a new hybrid operator, which creates a wealth for exploring feasible and promising individuals in the search space. We also direct other genetic operators in the correct way to produce a near-optimal final solution. To achieve tangible results, we have performed several scenarios. Compared to other existing approaches such as HEFT and QGARAR versions, our proposed algorithm has performed better in terms of average duration.