Earticle

The optimization of job-shop scheduling is very important because of its theoretical and practical significance. In this paper, a computationally effective approach of combining bacterial foraging strategy with particle swarm optimization for solving the minimum makespan problem of job shop scheduling is proposed. In the artificial bacterial foraging system, a novel chemotactic model is designed to address the job shop scheduling problem and a mechanism of quorum sensing and communication are presented to improve the foraging performance. In the particle swarm system, a novel concept for the distance and velocity of a particle is presented to pave the way for the job-shop scheduling problem. The proposed coevolutionary algorithm effectively exploits the capabilities of distributed and parallel computing of swarm intelligence approaches. The algorithm is examined using a set of benchmark instances with various sizes and levels of hardness and compared with other approaches reported in some existing literatures. The computational results validate the effectiveness of the proposed algorithm.

Complications of embedded applications are increasing. Within due time delivery to the market is also a pressure. So, the demand and use of real time operating systems is also increasing. However, it is an redundant fact that the Real Time Operating System (Real Time OS) can significantly degrade the performance. To face performance degradation, a Real-Time Task Manger (RTTM) has been presented in this paper. The Real-Time Task Manger is a hardware based extension to the processor that decreases performance bottleneck attributed to Real Time Operating System. Real-Time Task Manger decreases these performance bottlenecks introduced by Real Time Operating System by its hardware based architecture. The architecture of Real-Time Task Manger is being discussed in this paper provides an aid to deal with some common operations of Real Time Operating System that reason the performance degradation. For example event management, time management, and task scheduling. These operations have a property of some inborn parallelism. The Real-Time Task Manger uses this property to complete these operations in a constant time, and as a result, minimizes the overhead introduced by Real Time Operating System. The proposed architecture yields two benefits: It reduces the processor time taken by Real Time Operating System, and improves the response time to a considerable amount.

With the rapid progress of computational science and computer simulation ability, a lot of properties can be predicted by the powerful ability of parallel computation before the actual research and development. With the development of high performance computer architecture, GPU is more and more widely used in high performance computation field as an emerging architecture, and a growing number of computations use GPU heterogeneous cluster architecture. However, how to partition workload and map to computing resource has always been the focus and difficult point. In the current study of GPU, according to the problems of the computing power provided by each node and the cluster hardware architecture which the application programmers don't understand, some partitioning strategies will result in serious load imbalance problem. Aimed at the complexity brought by the different computing ability of the nodes of GPU clusters, this paper proposes a GPU data partitioning strategy of heterogeneous clusters based on learning. It collects the states of each node in the process of running a program, and then estimates the calculation ability of each node dynamically, so as to guide the data partitioning. Actual testing results show that, this strategy allocates different tasks to nodes based on computing ability to ensure load balancing among nodes, so as to improve the execution performance of CUDA programs on heterogeneous GPU clusters and it laid a solid foundation for efficient computing on heterogeneous GPU clusters.

Recent studies for network virtualization have shown a promising way to overcome the Internet ossification. The one of the key issues in network virtualization is a virtual network mapping problem, i.e., mapping a virtual network to the physical network. The situations of dynamic arrivals of virtual network request and the limited life cycle of the virtual networks pose significant challenges to the virtual network mapping problem. A balance between the resource allocation of the physical network and the number of mapped virtual networks. In this paper, we have considered the time characteristics that virtual network requests when mapping algorithms so as to achieve the objective that the node load and link load can simultaneously reach a balance. Giving full consideration to mutual restraints of time and resources, we propose a two-dimensional discrete weighted model based on time and resources, and establish a mathematical programming model of minimizing the degree of two-dimensional load balancing. Moreover, we devise a VN embedding algorithms LB-VNE. Simulation experiments show that the proposed algorithms can increase the acceptance ratio and the revenue by the substrate network in the long term.