Browsing by Author "Liu, Yen-Cheng"

Now showing 1 - 8 of 8

A branch-and-bound algorithm for makespan minimization in differentiation flow shops
(2013-12-01)
Liu, Yen-Cheng
;
Fang, Kuei-Tang
;
Lin, Bertrand
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
This article considers a differentiation flow-shop model, where the jobs are divided into various categories, each of which consists of two stages of operations. All products should be processed first on the single common machine at stage 1. At the second stage, each individual product proceeds to a dedicated machine according to its type. The problem of makespan minimization under the setting with two product types is known to be strongly NP hard. This article considers an arbitrary number of job types by developing a lower bound and two dominance rules, based upon which branch-and-bound algorithms are designed. Computational experiments are carried out to examine the performance of the proposed properties. The statistics show that the proposed properties can substantially reduce the computing efforts required for finding optimal solutions.
Detach and Adapt: Learning Cross-Domain Disentangled Deep Representation
(2018-01-01)
Liu, Yen-Cheng
;
Yeh, Yu-Ying
;
Fu, Tzu-Chien
;
Wang, Sheng-De
;
Chiu, Wei-Chen
While representation learning aims to derive interpretable features for describing visual data, representation disentanglement further results in such features so that particular image attributes can be identified and manipulated. However, one cannot easily address this task without observing ground truth annotation for the training data. To address this problem, we propose a novel deep learning model of Cross-Domain Representation Disentangler (CDRD). By observing fully annotated source-domain data and unlabeled target-domain data of interest, our model bridges the information across data domains and transfers the attribute information accordingly. Thus, cross-domain feature disentanglement and adaptation can be jointly performed. In the experiments, we provide qualitative results to verify our disentanglement capability. Moreover, we further confirm that our model can be applied for solving classification tasks of unsupervised domain adaptation, and performs favorably against state-of-the-art image disentanglement and translation methods.
Detection of Spatial-Modulated Signals in the Presence of CSI Error and Time-Spatial Correlation
(2013-01-01)
Chang, Hsuan-Cheng
;
Liu, Yen-Cheng
;
Su, Yu T.
;
傳播研究所
;
Institute of Communication Studies
Most of the studies on spatial modulation (SM) signal detection assume perfect channel state information (CSI) and ignore the channel\'s time-spatial correlation. Performance degradation incurs in the presence of imperfect CSI and channel correlations when estimating or predicting channel state based on periodic pilots. Detector structures that take into account both the channel estimation error and time-spatial correlation are investigated in this paper. We propose the CSI error and channel correlation aware ML detectors in conjunction with two channel estimation methods and present numerical results to quantify the associated performance improvements.
DOMAIN-ADAPTIVE GENERATIVE ADVERSARIAL NETWORKS FOR SKETCH-TO-PHOTO INVERSION
(2017-01-01)
Liu, Yen-Cheng
;
Chiu, Wei-Chen
;
Wang, Sheng-De
;
Wang, Yu-Chiang Frank
;
交大名義發表
Generating photo-realistic images from multiple style sketches is one of challenging tasks in image synthesis with important applications such as facial composite for suspects. While machine learning techniques have been applied for solving this problem, the requirement of collecting sketch and face photo image pairs would limit the use of the learned model for rendering sketches of different styles. In this paper, we propose a novel deep learning model of Domain-adaptive Generative Adversarial Networks (DA-GAN). The design of DA-GAN performs cross-style sketch-to-photo inversion, which mitigates the difference across input sketch styles without the need to collect a large number of sketch and face image pairs for training purposes. In experiments, we show that our method is able to produce satisfactory results as well as performing favorably against state-of-the-art approaches.
Initial Synchronization for Multi-Cell OFDMA Systems
(2009)
Liu, Yen-Cheng
;
Lin, Kuhn-Chang
;
Su, Yu T.
;
電信工程研究所
;
Institute of Communications Engineering
In this paper, a complete solution to the initial synchronization for multi-cell OFDMA (orthogonal frequency division multiple access) systems is proposed. The proposed method consists of the following operations: timing synchronization, fractional carrier frequency offset (CFO) estimation, fractional CFO compensation, and joint integer CFO and preamble index estimation. To estimate multiple fractional CFOs with pseudo-periodic preamble, we suggest an over-sampling approach. An ill-condition detection and noise suppression algorithm is also presented. For joint integer CFO and preamble index estimation, we use a differential detection based cross-correlation method. Simulation results have shown that the proposed method gives satisfactory performance for a wide SNR range even when the mobile station (MS) is on the cell boundaries.
New Constructions of Zero-Correlation Zone Sequences
(2013-08-01)
Liu, Yen-Cheng
;
Chen, Ching-Wei
;
Su, Yu T.
;
傳播研究所
;
Institute of Communication Studies
In this paper, we propose three classes of systematic approaches for constructing zero-correlation zone (ZCZ) sequence families. In most cases, these approaches are capable of generating sequence families that achieve the upper bounds on the family size (K )and the ZCZ width (T) for a given sequence period (N). Our approaches can produce various binary and polyphase ZCZ families with desired parameters (N, K, T) and alphabet size. They also provide additional tradeoffs amongst the above four system parameters and are less constrained by the alphabet size. Furthermore, the constructed families have nested-like property that can be either decomposed or combined to constitute smaller or larger ZCZ sequence sets. We make detailed comparisons with related works and present some extended properties. For each approach, we provide examples to numerically illustrate the proposed construction procedure.
On Composite Channel Estimation in Wireless Massive MIMO Systems
(2013-01-01)
Chen, Ko-Feng
;
Liu, Yen-Cheng
;
Su, Yu T.
;
傳播研究所
;
Institute of Communication Studies
A multiuser MIMO (MU-MIMO) TDD system in which the base station (BS) uses a large scale antenna array to communicate with single-antenna mobile users is considered. We present novel schemes to estimate both large-scale and small-scale fading coefficients by uplink pilot sequences. The large-scale fading coefficients (LSFCs) are usually ignored or assumed perfectly known and almost all MIMO channel state information (CSI) estimation studies deal solely with small-scale fading coefficients (SSFCs). We propose low complexity algorithms that give reliable estimates for both FCs. The LSFC estimator takes advantage of the excess spatial samples received and the channel hardening effect while the SSFC estimator is based on a rank-reduction approach. Our estimators require a low training overhead and both analysis and numerical experiments prove that they offer superior mean-squared error performance.
Systematic Constructions of Zero-Correlation Zone Sequences
(2009)
Chen, Ching-Wei
;
Liu, Yen-Cheng
;
Su, Yu T.
;
電信工程研究所
;
Institute of Communications Engineering
Sequences with desired autocorrelation (AC) and cross-correlation (CC) properties are often needed in many communication and radar system applications. In a multipath fading channel with a known delay spread bound, one needs a family of sequences that meets these desired correlation requirements only for those correlation lags that lie within a window called zero-correlation zone. This paper presents two classes of new systematic approaches a transform domain method and a direct (time domain) synthesis method for generating zero-correlation zone (ZCZ) sequences. The former approach generates sequence families that achieve the upper bounds on the family size and the zero-correlation zone width for a given sequence period. The latter unifies many existing construction methods and both provide more flexibilities in producing new families of sequences. We also suggest a modulation operation that is capable of transforming nonconstant modulus sequences into polyphase sequences.