Earticle

본 연구는 국가 AI 경쟁력을 정태적 순위 비교에 그치지 않고, 역량이 축적·확산되는 구조적 경 로와 정책 개입의 시간지연 효과를 규명하기 위해 시스템 다이내믹스(System Dynamics) 기반 분 석을 적용하였다. 분석 범위는 OECD 38개국이며, Tortoise Media의 Global AI Index가 제시하는 구 현(Implementation)·혁신(Innovation)·투자(Investment) 3대 축의 2023–2025년 점수(0–100 정규 화)를 활용하였다. 분석결과, OECD 내 점수 분포는 상위 소수 국가에 편중되어 있으며, 특히 투자 축에서 국가 간 격차가 크게 나타나 상위권 도약을 좌우하는 변별력 있는 축으로 기능하는 것으로 확인되었다. RQ1에 대해, 구현·혁신·투자는 ‘투자→혁신→채택(구현)→시장수요/상용화→추가투자’ 로 연결되는 강화루프와 ‘리스크 사건→신뢰 저하/규제·준수부담 증가→채택·투자 둔화’의 균형루프 가 동시에 작동하는 구조로 상호 영향을 주는 것으로 분석된다. RQ2에 대해, 정부 AI 예산은 R&D·공공 인프라를 통해 혁신(및 간접적으로 투자)을, 교육·훈련은 인재공급과 R&D 생산성을 통해 혁 신·구현을, 민간 AI 투자는 상용화·스케일업을 통해 투자 축을 가장 직접적으로 강화하며, 효과는 1.5–3년 내외의 시차를 두고 누적적으로 나타나는 것으로 분석되었다. RQ3에 대해, 한국이 투자 점수 OECD 3위에 도달하기 위해서는 민간 투자 레버리지의 상향을 핵심으로 하되 정부전략과 인재 기반을 결합한 패키지형 조합이 필요하며, 본 연구의 대표 조합(G=1.2, P=1.3, E=1.1)이 그 최소 조건 중 하나로 도출되었다.

This study applies a System Dynamics-based approach to move beyond static rank comparisons of national AI competitiveness and to identify the structural pathways through which capabilities accumulate and diffuse, as well as the time-lagged effects of policy interventions. The analysis covers 38 OECD countries and uses normalized (0-100) scores for 2023-2025 on the three pillars-Implementation, Innovation, and Investment-presented in Tortoise Media’s Global AI Index. The results show that score distributions within the OECD are concentrated among a small number of top countries. In particular, cross-country gaps are largest in the Investment pillar, confirming its role as a highly discriminating dimension that can determine whether a country breaks into the top tier. Regarding RQ1, Implementation, Innovation, and Investment are mutually reinforcing through a reinforcing loop-investment → innovation → adoption (implementation) → market demand/commercialization → additional investment-while simultaneously being constrained by a balancing loop in which risk events → declining trust / increased regulation and compliance burdens → slower adoption and investment. Regarding RQ2, government AI budgets strengthen Innovation (and indirectly Investment) through R&D and public infrastructure; education and training enhance Innovation and Implementation by expanding the talent supply and improving R&D productivity; and private AI investment most directly reinforces the Investment pillar via commercialization and scale-up. These effects accumulate over time and typically materialize with lags of approximately 1.5-3 years. Regarding RQ3, for Korea to reach third place in the OECD in the Investment score, increasing private-investment leverage should be the central lever, combined with a package of complementary measures that integrates government strategy and the talent base. One representative package identified in this study (G=1.2, P=1.3, E=1.1) is derived as one of the minimum conditions for achieving that target.

The structural dynamics in archwire-bending process is critical to the orthodontic archwire bending robot. In this paper, structure of orthodontic archwire bending robot is analyzed. Transient dynamic of orthodontic archwire bending is analyzed using the transient dynamic module of ANSYS/Workbench software. Stress curve and distribution of orthodontic archwire, fixed die, rotation die and supported end of orthodontic archwire in the process of orthodontic archwire bending is obtained. The structural dynamics simulation lays the foundation for the structural design of orthodontic archwire bending robot.

A series of aliphatic hydrocarbons, methane to hexane in the liquid state, are modeled with the molecular mechanical potential parameters treating all hydrogen degrees of freedom explicitly. Thermodynamic properties (heat capacities and heats of vaporization) are calculated from relatively short (20ps) molecular dynamics trajectories. The liquid state structures are also examined through various radial distribution functions. Molecular dynamics simulations reproduce experimentally measured properties within a few percent errors, thus indicate that the present set of all-hydrogen parameters is suitable for simulating macromolecular systems in bulk.

In this paper we have presented the results of thermodynamic, structural, and dynamic properties of model systems for liquid benzene, toluene and p-xylene in an isobaric-isothermal (NpT) ensemble at 283.15, 303.15, 323.15, and 343.15 K using molecular dynamics (MD) simulation. This work is initiated to compensate for our previous canonical (NVT) ensemble MD simulations [Bull. Kor. Chem. Soc. 2001, 23, 441] for the same systems in which the calculated pressures were too low. The calculated pressures in the NpT ensemble MD simulations are close to 1 atm and the volume of each system increases with increasing temperature. The first and second peaks in the center of mass g(r) diminish gradually and the minima increase as usual for the three liquids as the temperature increases. The three peaks of the site-site gC-C(r) at 283.15 K support the perpendicular structure of nearest neighbors in liquid benzene. Two self-diffusion coefficients of liquid benzene via the Einstein equation and via the Green-Kubo relation are in excellent agreement with the experimental measures. The self-diffusion coefficients of liquid toluene and p-xylene are in accord with the trend that the self-diffusion coefficient decreases with increasing number of methyl group. The friction constants calculated from the force auto-correlation (FAC) function with the assumption that the fast random force correlation ends at time which the FAC has the first negative value give a correct qualitative trends: decrease with increase of temperature and increase with the number of methyl group. The friction constants calculated from the FAC's are always less than those obtained from the friction-diffusion relation which reflects that the random FAC decays slower than the total FAC as described by Kubo [Rep. Prog. Phys. 1966, 29, 255].

Molecular dynamics simulations were performed to understand the structural arrangement of water molecules around highly charged nanoparticles under aqueous conditions. The effect of two highly charged nanoparticles on the solvation charge asymmetry has been examined. We calculated the radial distribution functions of the components of water molecules around nanoparticles which have four charge types at two different salt concentrations. Even though the distributions of water molecules surrounding a sodium ion and a chloride ion are hardly affected by the charges of nanoparticles and the salt concentrations, those around highly charged nanoparticles are strongly influenced by the charges of nanoparticles, but hardly by the charges of nanoparticles and salt concentrations. We find that the distributions of hydrogen atoms in water molecules around one highly charged nanoparticle are dependent on the magnitude of the nanoparticle charge.

FLOWERING LOCUS T (FT) and TERMINAL FLOWER 1 (TFL1) in Arabidopsis are homologous proteins that perform opposite functions: FT is an activator of flowering, and TFL1 is a repressor. It was shown before that change of a single amino acid (His88) of TFL1 to the corresponding amino acid (Tyr) of FT is enough to convert the floral repressor to an activator. However, structural basis of the functional conversion has not been understood. In our molecular dynamics simulations on modified TFL1 proteins, a hydrogen bond present in native TFL1 between the His88 residue and a residue (Asp144) in a neighboring external loop became broken by change of His88 to Tyr. This breakage induced conformational change of the external loop whose structure was previously reported to be another key functional determinant. These findings reveal that the two important factors determining the functional specificities of the floral regulators, the key amino acid (His88) and the external loop, are correlated, and the key amino acid determines the functional specificity indirectly by affecting the conformation of the external loop.

A polyelectrolyte chain confined in a slit nanochannel exhibits a structural transition from the one in free space. In this paper, the effect of the long-range electrostatic interactions between the xanthan polyelectrolyte and the slit wall on the confined xanthan conformation is investigated via the Brownian dynamics simulation. A neutral and two negatively charged surfaces of polydimethylsiloxane (PDMS) and glass are combined to make four kinds of slit channels with different charge characteristics: i) neutral-neutral, ii) glass-glass, iii) neutral-PDMS and iv) neutral-glass walls. Their walls are characterized by uniform surface charge densities determined from experimental data of zeta potential. Both the nonmonotonic chain size variation and the loss of long-range bond vector correlation, previously observed under confinement in the PDMS-PDMS slit, are also found in the neutral slit, demonstrating the nonelectrostatic origin of such crossover behaviors. As expected, the effect of wall charges is negligible at sufficiently high medium ionic strength of 100mM but it becomes significant in the opposite limit of 0.01mM. In the latter case, the high charge density of glass walls strengthens the effective confinement of a negatively charged polyelectrolyte and produces a xanthan structure comparable to that confined in a much narrower neutral slit. The obtained structural data suggest the possibility of controlling the structure of confined polyelectrolytes by the modification of surface charge characteristics of micro/nanofluidic devices in combination with the adjustment of the medium ionic strength.

부탄올의 구조이성질체(n-, iso-, sec-, tert-butanol) 와 n-부티르산에 대한 리파아제 효소.촉매 에스테르화 반응이 초임계 이산화탄소 조건 하에서 수행되었다. 본 실험은 교반속도 150 rpm, 반응 온도 323.15 K, 반응 압력 150 bar의 조건으로 고압반응기에서 5 h 동안 수행하였다. 실험에 사용된 리파아제는 Candida Antarctica lipase B (CALB)이다. 실험 결과는 HP-INNOWax 컬럼을 이용하여 FID (Flame Ionization detector)가 장착된 기체 크로마토그래피(Gas Chromatography, GC)를 이용하여 분석하였다. 반응 후 생성물의 전환률과 반응의 경향성을 분자동역학 시뮬레이션을 이용하여 예측된 결과와 정성적으로 비교하였다. 경쟁적인 저해반응이 포함된 Ping-Pong Bi-Bi 메커니즘을 기초로 하여, 반응의 각 단계를 적용하여 구조 최적화를 하였고 이를 이용해 전이상태의 에너지를 구하여 반응의 경향성을 예측하였다. 생성되는 에스테르 이성질체의 구조적 선호도는 분자동역학 시뮬레이션을 통하여 분석하였다. 이러한 방법의 개발은 앞으로 컴퓨터를 이용한 효소 반응의 설계에 유용하게 사용될 수 있을 것이다.

Lipase-catalyzed esterification of structural butanol isomers and n-butyric acid was investigated in supercritical carbon dioxide. The experiments were performed in a high pressure cell for 5 hrs with a stirring rate of 150 rpm at 323.15 K and 130 bar. The Candida Antarctica lipase B (CALB) was used in whole system as a catalyst. The experimental results were analyzed by GC-FID using a INNOWax capillary column. The conversion yield and the tendency of the esterification in supercritical carbon dioxide were compared with estimated results by molecular dynamics simulation. Based on the Ping-Pong Bi-Bi mechanism with competitive inhibition, each step of the reaction was optimized; using this result the transition state was predicted. Conformational preference of isomers was also analyzed using molecular dynamics simulations. This kind of approach will be further extended to the prediction of enzyme-catalyzed reactions using computers.