Narrow Results

The concept of aggregation within fuzzy sets has been deeply modified from the initial view as mappings where a fixed number of aggregated items, and the aggregation itself, take values in the unit interval to assume a family of mapping and impose certain restrictions to assure consistency and computability, till focus on aggregation as algorithms on lists of objects. This chapter points out how new concepts such as recursivity, map reducibility or hierarchy can be translated into computable aggregations.

As the energy consumption of storage systems is grows at a staggering rate, hybrid clusters have gained increasing importance as a potential approach to tackle this challenge. By introducing low-power nodes, data-driven companies like Facebook and Baidu have reduced the energy consumption effectively in their Master/Slave based storage systems. However, the Master/Slave based systems have several typical disadvantages such as low scalabilities and single points of failure. The P2P based systems with high scalabilities utilizes file location algorithms instead of table lookup mechanisms, thus resulting in a problem of how to utilize the different storage nodes discriminatively. In this paper, a hierarchical storage strategy called vnode hierarchical remapping (VHR) is proposed based on ’a P2P distributed system called ZDFS. The strategy guarantees the high scalability and viability of ZDFS, and takes advantage of different storage nodes. Several test cases running on X86 and ARM hybrid clusters are carried out, and the test results demonstrate that the VHR works well, it achieves a good I/O performance and low data access response time while reducing the energy consumption by 44.8%.

This study introduces an optimization method for solving the cross-coupled controller design of CNC machine. Based on the particle swarm optimization algorithm, wewi propose a hierarchical particle swarm optimization (HPSO) algorithm by adopting the concept of artificial bee colony algorithm, the instant update selection strategy which provides the newest information for other particles. Herein, we use the HPSO algorithm to solve the cross-coupled controller design and optimization of CNC machine. Simulation results and discussion are introduced to show the effectiveness and performance of the proposed approach.

Object-based storage is a kind of cloud storage technology, with better data sharing, better security, better intelligence. Mass storage system consists of thousands of devices, and gradually extended to form, therefore, the data placement is one of the key problems in the efficient organization of storage devices. Although consistent hashing has the good adaptability, but it need to increase the number of virtual nodes in heterogeneous environment. This paper introduces the hierarchical data placement algorithm, and improves the consistent hashing and improves the setting weight method and the clustering method.