• 스마트미디어저널
• /
• 제1권1호
• /
• pp.36-41
• /
• 2012

특징 선택은 패턴 인식의 성능을 향상시키기 위해 부분집합을 구성하는 중요한 문제다. 특징 선택에는 순차 탐색 알고리즘으로부터 확률 기반의 유전 알고리즘까지 다양한 접근 방법이 적용 되었다. 본 연구에서는 특징 선택을 위해 양자 비트, 상태의 중첩 등 양자 컴퓨터 개념을 기반으로 하는 양자 기반 유전 알고리즘(QGA: Quantum-inspired Genetic Algorithm)을 적용하였다. QGA 성능은 전통적인 유전 알고리즘(CGA: Conventional Genetic Algorithm)을 적용한 특징 선택 방법과 분류율 및 평균 특징 개수의 비교를 통해 이루어졌으며, UCI 데이터를 이용한 실험 결과 QGA를 적용한 특징 선택 방법이 CGA를 적용한 경우에 비해 전반적으로 좋은 성능을 보임을 확인 할 수 있었다.

• 한국환경과학회지
• /
• 제18권3호
• /
• pp.299-308
• /
• 2009

We manufactured PVA-derived hydrogels using a foam generation technique that has been widely used to prepare colloidal gas aphrons(CGA). These gels were differentiated to the conventional gels such as for medical or pharmaceutical applications, which have tiny pores and some crystalline structure. Rather these should be used in de-pollution devices or adhesion of cells or biomolecules. The crosslinkers used in this work were amino acid, organic acid, sugars and lipids(vitamins). The structures of the gels were observed in a scanned electron microscope. Amino acids gels showed remarkably higher swelling ratios probably because their typical functional groups help constructing a highly crosslinked network along with hydrogen bonds. Boric acid and starch would catalyze dehydration while structuring to result in much lower water content and accordingly high gel content, leading to less elastic, hard gels. Bulky materials such as ascorbic acid or starch produced, in general, large pores in the matrices and also nicotinamide, having large hydrophobic patches was likely to enlarge pore size of its gels as well since the hydrophobicity would expel water molecules, thus leading to reduced swelling. Hydrophilicity(or hydrophobicity), functional groups which are involved in the reaction or physical linkage, and bulkiness of crosslinkers were found to be more critical to gel's cross linking structure and its density than molecular weights that seemed to be closely related to pore sizes. Microscopic observation revealed that pores were more or less homogeneous and their average sizes were $20{\mu}m$ for methionine, $10-15{\mu}m$ for citric acid, $50-70{\mu}m$ for L-ascorbic acid, $30-40{\mu}m$ for nicotinamide, and $70-80{\mu}m$ for starch. Also a sensory test showed that amino acid and glucose gels were more elastic meanwhile acid and nicotinamide gels turned out to be brittle or non-elastic at their high concentrations. The elasticity of a gel was reasonably correlated with its water content or swelling ratio. In addition, the PVA gel including 20% ascorbic acid showed fair ability of cell adherence as 0.257mg/g-hydrogel and completely degraded phenanthrene(10 mM) in 240 h.

• Structural Engineering and Mechanics
• /
• 제16권5호
• /
• pp.557-580
• /
• 2003

Laminated composite structures find wide range of applications in many branches of technology. They are much suited for weight sensitive structures (like aircraft) where thinner and lighter members made of advanced fiber reinforced composite materials are used. The orientations of fiber direction in layers and number of layers and the thickness of the layers as well as material of composites play a major role in determining the strength and stiffness. Thus the basic design problem is to determine the optimum stacking sequence in terms of laminate thickness, material and fiber orientation. In this paper, a new optimization technique called Cellular Automata (CA) has been combined with Genetic Algorithm (GA) to develop a different search and optimization algorithm, known as Cellular Genetic Algorithm (CGA), which considers the laminate thickness, angle of fiber orientation and the fiber material as discrete variables. This CGA has been successfully applied to obtain the optimal fiber orientation, thickness and material lay-up for multi-layered composite hybrid beams plates and shells subjected to static buckling and dynamic constraints.