Search

An efficient scheduling of dependent-tasks over edge-cloud is the key for achieving better utilization of computational resources and timely completion of interdependent tasks for scientific as well as defence-oriented applications. Many applications comprise several tasks that are dependent in nature and are required to be executed in a specific sequence within a minimum span of time. To handle the execution of such dependent-tasks on distributed computational resources is a challenging problem; as the number of tasks increase, the solution space comprising different task-resource mapping sequences also increase exponentially and it is very difficult to find the near optimal solutions in the search space. In this paper, we focus on two strategies for obtaining optimal solutions for scheduling the multiple dependent-tasks with the specified sequence in a parallel and distributed environment. In the first approach, a hybrid mechanism is proposed to efficiently search the scheduling solution space for multiple dependent tasks. The idea is to first find a schedule by heuristic algorithms and use these as initial solutions in the search space to obtain better solutions using unsupervised machine learning methods. In the second approach, each task-graph is partitioned into different clusters of sub-tasks, where each partitioned cluster is mapped onto the same resource. This strategy reduces the idle times on the resources, wherever possible, in comparison with the first approach. The innovative part of the proposed approaches is to schedule the multiple dependent-tasks in a parallel fashion rather than scheduling in sequential manner. The results show that the schedules obtained by our proposed approaches minimize the total execution time (TET) significantly as compared to other approaches considered in our research study. [ABSTRACT FROM AUTHOR]

Traditional algorithms have been used in determining solutions for the grid scheduling problem. However, it has been witnessed in the recent times that the increased complexity and size of the job have been a reason for the degradation in algorithm's performance. This led to the emergence of a new class of heuristic algorithms that gives optimal or near-optimal solutions to optimization problems. The proposed algorithm solves multi-objective grid scheduling problem and performs further optimization. The proposed swarm-based algorithm is a threshold-constrained Ant Colony Optimization using Shortest Job First (SJF) in initialization step. The characteristics of proposed algorithm are as follows. It minimizes the job completion (execution) time and enhances resource utilization. It supports load balancing and scalability. The performance test of algorithms is carried out using a real-time dataset, which can set benchmark results for future research. Simulation results reveal that the proposed swarm algorithm obtains promising results over other algorithms taken for comparison. [ABSTRACT FROM AUTHOR]

As an emerging novel evolutionary technique, the immune algorithm has gained a lot of attention and wide applications in various fields of engineering and management. To our knowledge, this paper first considers the application of immune algorithm for the classic permutation flow shop scheduling problem. We present a hybrid heuristic combining immune algorithm and simulated annealing for the n-job, m-machine permutation flow shop scheduling problem to minimize makespan and total flow time. A heuristically directed population-based construction approach in the immune algorithm and a forward/backward shift neighborhood search heuristic in the simulated annealing of the proposed method is utilized to explore the search process for better schedules of jobs as well to speed up the convergence speed of the proposed algorithm. The proposed method is tested with Taillard's flow shop scheduling benchmark instances for different problems with job sizes varying from 20 to 500. The computational results demonstrate that the proposed heuristic is very competitive with the state-of-the-art procedures in terms of both solution quality and computational times. [ABSTRACT FROM AUTHOR]

For heterogeneous distributed computing systems, important design issues are scalability and system optimization. Given such systems, it is crucial to develop low computational complexity algorithms to schedule tasks in a manner that exploits the heterogeneity of the resources and applications. In this paper, we report and evaluate three scalable, and fast scheduling heuristics for highly heterogeneous distributed computing systems. We conduct a comprehensive performance evaluation study using simulation. The benchmarking outlines the performance of the schedulers, representing scalability, makespan, flowtime, computational complexity, and memory utilization. The set of experimental results shows that our heuristics perform as good as the traditional approaches, for makespan and flowtime, while featuring lower complexity, lower running time, and lower used memory. The experimental results also detail the various scenarios under which certain algorithms excel and fail. [ABSTRACT FROM AUTHOR]

This paper presents an efficient heuristic method to minimize total flow time in no-wait flow shop scheduling. It is based on the assumption that the priority of a job in the initial sequence is given by the sum of its processing times on the bottleneck machines. Empirical results demonstrate the superiority of the proposed method over the best-known heuristics in the literature, while remaining the same complexity order of O( n2). [ABSTRACT FROM AUTHOR]