This study develops and evaluates a prompt-driven large language model (LLM) agent for section design of doubly reinforced concrete (RC) beams. Using Google Gemini (Gems), an engineering “expert” that operates without fine-tuning by uploading ACI-318 provisions, sample design documents, and a database of prior beam designs was developed. The agent interprets code clauses, formulas, and constraints from these materials and retrieves similar design cases to propose an initial solution. It then incorporates user-specified natural-language constraints—most notably a strength-ratio cap (design strength ≤ 105% of required strength)—to iteratively refine toward safe and economical designs. Beyond reporting member size and reinforcement details, the agent provides step-by-step computational justifications for moment and shear checks, increasing verifiability and instructional value. We benchmark the LLM-generated designs against results from the commercial program MIDAS/Design+ and observe close agreement. In several scenarios, the constraint-guided LLM solutions are more material-efficient while remaining code-compliant. The workflow also supports batch processing from spreadsheet inputs, enabling practical automation across multiple beams. The approach requires no additional model training or coding making it accessible to non-developer practitioners. Results indicate that a general-purpose LLM, properly grounded with code documents and examples, can achieve practice-level performance with transparent reasoning. This demonstrates a viable approach to AI-assisted structural design that is explainable, interactive, and readily integrated with engineering workflows.

This study evaluated the field-scale performance of an amorphous iron hydroxide (Fe(OH)3)-based desulfurizing agent for the removal of sulfur-based odorous compounds emitted from wastewater treatment facilities, including equalization tanks and sludge dewatering unit facilities. Hydrogen sulfide (H2S), methyl mercaptan (MM), dimethyl sulfide (DMS), and dimethyl disulfide (DMDS), which account for over 60~80% of total odor impact in such facilities, were targeted in this research. A drytype adsorption system packed with porous amorphous Fe(OH)3 was installed at a wastewater treatment plant and operated continuously for 45 days. Odorous gas concentrations were measured before and after treatment using portable analyzers and gas chromatography-pulsed flame photometric detector (GC-PFPD). The desulfurizing agent demonstrated a high H2S removal efficiency of over 99.9%, even under high inlet concentrations exceeding 500 ppm. Physicochemical analyses including XRD, XRF, EDS and BET confirmed that the material was amorphous, possessed a high surface area (243.4 m2/g), and exhibited a mesoporous structure favorable for gas adsorption. Hysteresis observed in nitrogen adsorption isotherms indicated a bottleneck-shaped pore structure, which enhances adsorption of odorous gases and removal efficiency. Notably, the system maintained stable performance under varying humidity without significant degradation.

As a key component of composite materials, the interface quality is crucial for determining the mechanical properties of composites. Carbon fiber sizing treatment significantly enhances the fiber-matrix interface, a process extensively utilized in the carbon fiber industry. This study synthesized an environmentally friendly waterborne polyurethane sizing agent and investigated the impact of molecular weight, a critical factor, on composite performance by varying the soft segment type in the polyurethane. This research provides insights into cost-effective and eco-friendly surface treatment methods for carbon fibers and the design of robust interface structures.

The emergence and re-emergence of infectious diseases pose ongoing threats to public health. This study aims to develop an agent-based simulation model (ABM) to predict the spread of novel infectious diseases during early outbreak phases and evaluate the effectiveness of control measures, specifically focusing on the impact of interventions such as maskwearing, vaccination, and social distancing on outbreak dynamics and the reduction of symptomatic cases. Using demographic and COVID-19 outbreak data from South Korea, we constructed a detailed contact network model encompassing workplaces, schools, households, and communities. Using demographic and COVID-19 outbreak data from Seoul, South Korea, we constructed a detailed contact network model encompassing workplaces, schools, households, and communities. Key transmission parameters were inferred using Approximate Bayesian Computation. The resulting ABM platform, implemented in a C-based R package, allows for flexible scenario simulation involving 56 adjustable parameters, including mask-wearing, vaccination coverage, and social distancing. Simulation outputs demonstrated the model’s capacity to reproduce observed transmission patterns in workplace and school outbreaks, enabling public health authorities to anticipate outbreak dynamics and assess interventions. This framework provides a valuable decision-support tool for controlling future infectious disease incursions.

Reinforcement learning (RL) is successfully applied to various engineering fields. RL is generally used for structural control cases to develop the control algorithms. On the other hand, a machine learning (ML) is adopted in various research to make automated structural design model for reinforced concrete (RC) beam members. In this case, ML models are developed to produce results that are as similar to those of training data as possible. The ML model developed in this way is difficult to produce better results than the training data. However, in reinforcement learning, an agent learns to make decisions by interacting with an environment. Therefore, the RL agent can find better design solution than the training data. In the structural design process (environment), the action of RL agent represent design variables of RC beam. Because the number of design variables of RC beam section is many, multi-agent DQN (Deep Q-Network) was used in this study to effectively find the optimal design solution. Among various versions of DQN, Double Q-Learning (DDQN) that not only improves accuracy in estimating the action-values but also improves the policy learned was used in this study. American Concrete Institute (318) was selected as the design codes for optimal structural design of RC beam and it was used to train the RL model without any hand-labeled dataset. Six agents of DDQN provides actions for beam with, beam depth, bottom rebar size, number of bottom rebar, top rebar size, and shear stirrup size, respectively. Six agents of DDQN were trained for 5,000 episodes and the performance of the multi-agent of DDQN was evaluated with 100 test design cases that is not used for training. Based on this study, it can be seen that the multi-agent RL algorithm can provide successfully structural design results of doubly reinforced beam.

As digital transformation accelerates, platform business has become a core business model in modern society. Platform business has a network effect where the winner takes all. For this reason, it is crucial for a company's pricing policy to attract as many customers as possible in the early stages of business. Telecommunication service companies are experiencing stagnant growth due to the saturation of the smartphone market and intensifying competition in rates, but the burden of maintaining communication networks is increasing due to the rapid increase in traffic caused by domestic and foreign CSPs. This study aims to understand the dynamic characteristics of the telecommunications market by focusing on pricing policy. To this end, we analyzed how ISPs, CSPs, and consumers react to changes in pricing policy based on the prisoner's dilemma theory. The analysis of the dynamic characteristics of the market was conducted through simulation using the Agent-Based Model.

The surface treatment processes of carbon fibers is very important, because of their significant impact on fiber handling, filament protection, and interfacial properties. In this study, the effects of two different sizing agents with different molecular weights, with or without a nonionic surfactant, on the performance of a melt-spun polyacrylonitrile-based carbon fiber and carbon fiber/epoxy interfacial adhesion are investigated. The focusing property and spread-ability of a low-molecularweight sizing agent with a surfactant show outstanding performances because of the high penetration between the fibers and high interfacial bonding with the fibers. In addition, wettability of the matrix (epoxy resin) of the low-molecular-weight sizing agent are superior to those of the high-molecular-weight sizing agent. Furthermore, the nonionic surfactant used as an assistant improves the sizing amount and wettability by forming micelles with the epoxy. The interfacial shear strength (IFSS) of the low-molecular-weight sizing agent with a surfactant is also superior to that of other sizing agents. The IFSS is closely related to the sizing amount of the coating on the carbon fiber surface and matrix wettability.

선박에는 단열을 위한 발포제가 적용된다. 기존의 발포제에는 지구온난화물질인 수소불화탄소(HFC)를 다량 포함하고 있는 문제점이 있으며, 우리나라는 몬트리올 의정서의 ‘키칼리 개정서’를 채택함에 따라 HFC를 ‘24년부터 ’45년까지 기준 수량의 80% 감 축하기로 결정되었다. 이에, 메틸포메이트 원료는 지구온난화지수가 0(HFC는 960~1,430)으로 향후 친환경발포제로 높은 기대를 갖고 있다. 하지만, 메틸포메이트 발포제의 성능은 원료의 순도 및 주변환경에 높은 영향을 받음으로 각 공정환경에 대한 정확한 분류가 필요하다. 이에, 본 논문에서는 주변환경(온도)과 메틸포메이트 순도에 따라, 총 4개의 케이스를 만들었다. 각 케이스에 대해서 10,010 장의 이미지를 학습하고, 이를 구글넷(GoogLeNet)알고리즘을 이용하여 분류하였다. 분류결과 정확도는 96.8%를 갖고, F1-Score는 0.969 를 갖는 것으로 계산하였다.

As digital transformation accelerates, platform business has become a core business model in modern society. Platform business has a network effect where the winner takes all. For this reason, it is crucial for a company's pricing policy to attract as many customers as possible in the early stages of business. Telecommunication service companies are experiencing stagnant growth due to the saturation of the smartphone market and intensifying competition in rates, but the burden of maintaining communication networks is increasing due to the rapid increase in traffic caused by domestic and foreign CSPs. This study aims to understand the dynamic characteristics of the telecommunications market by focusing on pricing policy. To this end, we analyzed how ISPs, CSPs, and consumers react to changes in pricing policy based on the prisoner's dilemma theory. The analysis of the dynamic characteristics of the market was conducted through simulation using the Agent-Based Model.

MES(manufacturing execution system) plays a critical role in improving production efficiency by managing operations across the entire manufacturing system. Conventional manufacturing systems employ a centralized control structure, which has limitations in terms of the flexibility, scalability and reconfigurability of the manufacturing system. Agent-based manufacturing systems, on the other hand, are better suited to dynamic environments due to their inherent high autonomy and reconfigurability. In this study, we propose an agent-based MES and present its collaboration model between agents along with a data structure. The agent-based MES consists of three types of core agents: WIPAgent, PAgent(processing agent), and MHAgent(material handling agent). The entire manufacturing execution process operates through collaboration among these core agents, and all collaboration is carried out through autonomous interactions between the agents. In particular, the order-by-order dispatching process and the WIP(work-in-process) routing process are represented as respective collaboration models to facilitate understanding and analyzing the processes. In addition, we define data specifications required for MES implementation and operation, and their respective structures and relationships. Moreover, we build a prototype system employing a simulation model of an exemplary shop-floor as a simulation test bed. The framework proposed in this study can be used as a basis for building an automated operating system in a distributed environment.

Recently, in the manufacturing industry, changes in various environmental conditions and constraints appear rapidly. At this time, a dispatching system that allocates work to resources at an appropriate time plays an important role in improving the speed or quality of production. In general, a rule-based static dispatching method has been widely used. However, this static approach to a dynamic production environment with uncertainty leads to several challenges, including decreased productivity, delayed delivery, and lower operating rates, etc. Therefore, a dynamic dispatching method is needed to address these challenges. This study aims to develop a reinforcement learning-based dynamic dispatching system, in which dispatching agents learn optimal dispatching rules for given environmental states. The state space represents various information such as WIP(work-in-process) and inventory levels, order status, machine status, and process status. A dispatching agent selects an optimal dispatching rule that considers multiple objectives of minimizing total tardiness and minimizing the number of setups at the same time. In particular, this study targets a multi-area manufacturing system consisting of a flow-shop area and a cellular-shop area. Thus, in addition to the dispatching agent that manages inputs to the flow-shop, a dispatching agent that manages transfers from the flow-shop to the cellular-shop is also developed. These two agents interact closely with each other. In this study, an agent-based dispatching system is developed and the performance is verified by comparing the system proposed in this study with the existing static dispatching method.

In this study, fire extinguisher system to which form fire extinguisher agents were adopted was applied to the combat vehicle crew room to apply fire extinguishing performance and acid gas safety that meet the national defense standards. As a result of evaluation and verification, the following conclusions were drawn. For standard fire sizes in the combat vehicle crew's standard model, we ignited using a mixture of Novec 1230 and Halon 1301 form extinguisher agent and released form extinguisher agent after 30 seconds to determine the fire extinguishing time. The amount of acid gas generated met the criteria in all cases. When the fire size was increased to 0.12m2 and a 2.0mm nozzle was used, all of the extinguishing time, the amount of acid gas generated, and the concentration of Novec 1230 met the criteria. Despite the more difficult conditions to extinguish the fire by making the fire larger, it was possible to confirm the extinguishing performance of the Novec 1230 form extinguisher agent and its safety against acid gas.

Lithium-ion batteries are widely used in various advanced devices, including electric vehicles and energy storage devices. As the application range of lithium-ion batteries expands, it will be increasingly important to improve their gravimetric and volumetric energy density. Layer-structured oxide materials have been widely adopted as cathode materials in Li-ion batteries. Among them, LiNiO2 has attracted interest because of its high theoretical capacity, ~274 mAh g-1, assuming reversible one Li+-(de)intercalation from the structure. Presently, such layered structure cathode materials are prepared by calcination of precursors. The precursors are typically hydroxides synthesized by coprecipitation reaction. Precursors synthesized by coprecipitation reaction have a spherical morphology with a size larger than 10 μm. Spherical precursors in the several micrometer range are difficult to obtain due to the limited coprecipitation reaction time, and can lead to vigorous collisions between the precursor particles. In this study, spherical and small-sized Ni(OH)2 precursors were synthesized using a new synthesis method instead of the conventional precipitation method. The highest capacity, 170 mAh g-1, could be achieved in the temperature range of 730~760 °C. The improved capacity was confirmed to be due to the higher quality of the layered structure.

모빌리티 예측은 단순한 통행 경로 예측을 넘어, 사회 전반의 효율성 및 안전성 향상을 위한 핵심 데이터를 제공한다는 점에서 중 요하다. 기존의 예측 기법은 시공간적 규칙성과 개인 이동 패턴의 통계적 특성 분석에 주로 의존하였으며, 최근 딥러닝 기반의 시공간 모델링을 통해 예측 성능이 향상되었다. 그러나 여전히 개인 통행의 단기·장기적 시공간 의존성 및 복잡한 패턴을 처리하는 데 한계가 존재한다. 이를 극복하기 위해, 본 연구는 대규모 사전 학습된 거대 언어 모델(Large Language Model; LLM)을 도입하여, 개인 속성뿐 만 아니라 실제 통행 데이터를 반영한 객체 단위 통행 생성 프레임워크를 제안한다. LLM 기반(ChatGPT-4o) 객체 단위 통행 생성 프레 임워크는 (1) 개인 모빌리티 패턴 학습, (2) 통행 생성의 두 단계로 이루어진다. 이후 한국교통연구원의 개인통행 실태조사(2021) 데이 터를 이용하여 프레임워크의 통행 생성 성능을 확인하였다. 통행 시작·출발 시간 분포, 출발·도착지 장소 유형, 통행목적, 이용 교통수 단의 정확도를 확인한 결과, 대부분 항목에서 70% 이상의 정확도를 보였다. 하지만 통행목적은 13개의 목적 중 하나를 예측해야 하기 에 정확도가 다른 항목에 비해 약 40%로 낮게 나타났다. 본 연구는 통행 생성 프레임워크를 설계하고, 이에 맞춰 입력 데이터를 가공 및 프롬프트 엔지니어링을 수행함으로써 LLM 기반 통행 생성 기술의 가능성을 확인하였다. 향후 프레임워크의 예측 성능 검증 및 개 선을 위한 추가 연구가 필요하며, 날씨, 대규모 행사 등과 같은 외부 요인들을 고려하면 더욱 정교하고 현실적인 통행일지를 생성할 수 있을 것이다.