Browsing by Author "Shyu, Shyong Jian"

Now showing 1 - 4 of 4

Discrete particle swarm optimization for materials acquisition in multi-unit libraries
(2012)
Wu, Yi-Ling
;
Shyu, Shyong Jian
;
Ho, Tsu-Feng
;
Lin, Bertrand M. T.
;
資訊管理與財務金融系註：原資管所+財金所
With limited budget and inflation rate, the materials acquisition in multi-unit libraries has been a challenging issue all over the world. The materials acquisition for multi-unit libraries can be regarded as a generalized version of the knapsack problem, which was known to be NP-hard, with much more constraints. Thus, it can be computationally expensive to solve the problem. In this paper, the materials acquisition problem in multi-unit libraries is formulated as an integer programming model, and two different constraint-handling mechanisms applied in discrete particle swarm optimization algorithm for obtaining the near optimal solution are presented. It is evident from our computational results that one constraint-handling mechanism can solve the problem effectively and efficiently, while the other one takes more time.
Discrete Particle Swarm Optimization with Scout Particles for Library Materials Acquisition
(2013)
Wu, Yi-Ling
;
Ho, Tsu-Feng
;
Shyu, Shyong Jian
;
Lin, Bertrand M. T.
;
資訊管理與財務金融系註：原資管所+財金所
Materials acquisition is one of the critical challenges faced by academic libraries. This paper presents an integer programming model of the studied problem by considering how to select materials in order to maximize the average preference and the budget execution rate under some practical restrictions including departmental budget, limitation of the number of materials in each category and each language. To tackle the constrained problem, we propose a discrete particle swarm optimization (DPSO) with scout particles, where each particle, represented as a binary matrix, corresponds to a candidate solution to the problem. An initialization algorithm and a penalty function are designed to cope with the constraints, and the scout particles are employed to enhance the exploration within the solution space. To demonstrate the effectiveness and efficiency of the proposed DPSO, a series of computational experiments are designed and conducted. The results are statistically analyzed, and it is evinced that the proposed DPSO is an effective approach for the studied problem.
Material Acquisitions in Academic Libraries
(2008)
Ho, Tsu-Feng
;
Shyu, Shyong Jian
;
Wu, Yi-Ling
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
As the population of books and other materials increases explosively, an effective plan for material acquisitions becomes an emergent and significant need for most modern academic libraries. However there are few researches on this topic in the past decades. We thereby propose a formal model for material acquisitions that resembles the practical situations of academic libraries. In the proposed model, we consider several features for each material (including the preference, price and its categories) as well as the budget for each category. The goal is to select materials to achieve a maximum total preference under the constraint that the acquisition for each category of materials is controlled by a predefined budget. With the computational complexity of the studied problem, we seek to produce approximate solution in a reasonable time. Three meta-heuristics, namely simulated annealing, genetic algorithm and tabu search, are developed to cope with this problem. Computational results reveal that tabu search is the most elegant approach among the three for the material acquisitions problem.
Secret image sharing with smaller shadow sizes for general access structures
(2016-11)
Chen, Ying-Ru
;
Chen, Ling-Hwei
;
Shyu, Shyong Jian
;
資訊工程學系
;
Department of Computer Science
In the area of secret image sharing (SIS), most papers focused on the schemes for threshold or some special access structures. Regarding general access structures (GAS), few results have been found in the literature. Two SIS schemes for GAS were proposed in 2001 and 2010, both are based on qualified sets. However, one distorts the reconstructed secret image, and some extra information is needed in both schemes. Here, we propose three polynomial based SIS schemes for GAS. Considering either qualified or forbidden sets, these schemes can reconstruct the secret image perfectly without any extra information needed. Some proof and analysis on the shadow sizes of the three schemes are given to lead us to choose the one with the smallest size. In addition, we also give some comparisons with two existing schemes, and security issue is also addressed in conclusion.