Hsun-Wen Chang

The Structural Birnbaum Importance of Consecutive-k Systems

This paper is concerned with three types of structural Birnbaum importance of components in a consecutive-k system. We accomplish the following three tasks:(i) Survey existing results, including pointing out results claimed but proofs are incomplete.(ii) Give some new results and useful lemmas.(iii) Observe what are still provable and what are not with the help of extensive computing.

On extensibilities of interconnection networks

Researchers have used number of nodes to measure the extensibility of a topology. However, this metric is not very evident. In this paper, we introduce a specific metric called extensible density to measure the extensibilities of interconnection networks. Some topologies have high degree of extensibilities, but efficient parallel algorithms can apply only on a special subclass of these topologies. Furthermore, we extend the concept of density to measure the applicable extent of parallel algorithms.

Fault-tolerant routing in unique-path multistage Omega network

Abstract In this paper, two new near one-pass packet-routing schemes are proposed to tolerate all single faults among the switching boxes or the connecting links respectively in a unique-path multistage Omega network. Working in an undirected network, our results can compete with the two-pass routing scheme, which works in a directed network, proposed by Varma.

NOVEL PIPELINING AND PROCESSOR ALLOCATION STRATEGY FOR MONOID COMPUTATIONS ON UNSHUFFLE-EXCHANGE NETWORKS

This short paper presents a novel pipelining and processor allocation strategy for monoid computations on an unshuffle-exchange network. In the strategy, the processor utilization is near 1 and the communication is collision-free. With the characteristics of constant connections to each processor and only a single output node on the network, the method given here can compete with the method of Barnard and Skillicorn based on a hypercube network with multiple output nodes.

Enumeration of RNA Hairpins and Cloverleaves

Abstract RNA (Ribonucleic acid) is an important molecule in the cell, which participates in a lot of basic biological functions. Hairpins and cloverleaves are two important structures in RNA secondary structures. In this paper, we use generating functions to enumerate RNA hairpins and Cloverleaves. We also point out incorrect results in literature.

The 3-Wide-Diameter of the Hyper-L Triple-Loop Network

The wide-diameter of a network has been considered as an important measurement of fault-tolerance. The 3-wide-diameter of a network is defined as the minimum length l such that, between any two nodes, there are three disjoint paths of length at most l. With small diameter D and simple structure, the hyper-L triple-loop network, denoted by HL(l, m, n), has been showed to be a good construction for a triple-loop network. Hwang and Lin proved that the 3-wide-diameter of hyper-L triple-loop network is at most D + 2. In this paper, we exactly present the S-wide-diameter of hyper-L triple-loop network. We prove that the 3-wide-diameter is D + 1 except two cases with D + 2

New families of triple-loop networks

There are two important topics in the multi-loop network. One is to fmd the minimum diameter for a fixed number of nodes, which is corresponding with the problem of minimizing the transmission delay. The other is to fmd the maximum number of nodes for a fixed diameter, which is corresponding with the problem of maximizing the support of services. For most triple-loop networks, however, the sharp bounds have not been discovered. In this paper, we propose five new families of triple-loop networks to improve results of these two topics.

Optimizing one-to-one data communication on double-loop networks

A double-loop network DL(n; a, b) has n nodes 0, 1, ..., n-1 and 2n links u rarr u+a (mod n) and u rarr u+b (mod n) where u is a node and n, a, and b are positive integers. Since DL(n; a, b) is 2-regular, there are at most two disjoint paths. In this paper, we will optimize one-to-one data communication with the store-arid-forward model and in the packet-switching mode on the double-loop network. We first find two disjoint paths and then partition workloads according to their lengths such that the transferring time is minimized