Search

Purpose - In many decision-making problems under parameter uncertainty, the most popular stochastic approach is not used because of its serious drawbacks. The purpose of this paper is to present another approach, which copes with the uncertainty of parameters. It uses a precise criterion evaluating a decision with respect to uncertain parameters. This precision by the maximum operator is performed on a term based on the criterion and called the relative regret. The approach is applied to the allocation problems in a complex of operations. Design/methodology/approach - The resource allocation problems in a complex of operations of independent and dependent structures to minimize a total execution time of all operations are investigated. Then, the results are extended for the problem of a task allocation in the complex of independent operations. The case is considered when the parameters in the functional models of the operations are uncertain, and their values belong to the intervals of known bounds. The solution algorithms for the uncertain problems are based on known solution algorithms for the corresponding deterministic problems. The solution algorithms for the latter problems are outlined in the paper. Findings - The main contribution of the paper consists in presenting the property that it is possible for the uncertain problems considered to replace the solution of the uncertain allocation problems by solving a number of corresponding deterministic problems. Research limitations/implications - The useful and interesting property of the solution algorithm for the allocation problems, in general, cannot be applied to the other decision-making problems under uncertainty. As an example of such a problem, a simple routing-scheduling problem is presented for which, however, a number of possible parameter scenarios can be substantially limited. Practical implications - The allocation problems addressed in the paper have a variety of applications in computer systems and in manufacturing systems. Moreover, a lack of crisp values for the parameters in models of individual operations is rather common. Originality/value - The paper extends previous results for the allocation problems in a complex of operations. [ABSTRACT FROM AUTHOR]

Cloud computing represents a novel computing model in the contemporary technology world. In a cloud system, the computing power of virtual machines (VMs) and network status can greatly affect the completion time of data intensive tasks. However, most of the current resource allocation policies focus only on network conditions and physical hosts. And the computing power of VMs is largely ignored. This paper proposes a comprehensive resource allocation policy which consists of a data intensive task scheduling algorithm that takes account of computing power of VMs and a VM allocation policy that considers bandwidth between storage nodes and hosts. The VM allocation policy includes VM placement and VM migration algorithms. Related simulations show that the proposed algorithms can greatly reduce the task completion time and keep good load balance of physical hosts at the same time. [ABSTRACT FROM AUTHOR]

In this paper, we consider a class of stochastic resource allocation problems where resources assigned to a task may fail probabilistically to complete assigned tasks. Failures to complete a task are observed before new resource allocations are selected. The resulting temporal resource allocation problem is a stochastic control problem, with a discrete state space and control space that grow in cardinality exponentially with the number of tasks. We modify this optimal control problem by expanding the admissible control space, and show that the resulting control problem can be solved exactly by efficient algorithms in time that grows nearly linear with the number of tasks. The approximate control problem also provides a bound on the achievable performance for the original control problem. The approximation is used as part of a model predictive control (MPC) algorithm to generate resource allocations over time in response to information on task completion status. We show in computational experiments that, for single resource class problems, the resulting MPC algorithm achieves nearly the same performance as the optimal dynamic programming algorithm while reducing computation time by over four orders of magnitude. In multiple resource class experiments involving 1000 tasks, the model predictive control performance is within 4% of the performance bound obtained by the solution of the expanded control space problem. [ABSTRACT FROM AUTHOR]