인공지능 기술이 발전함에 따라, 과거의 범죄 관련 데이터를 광범위하게 수집ㆍ분석하여 미래의 범죄발생 시간ㆍ장소나 범죄자, 피해자를 예측 하고 특정인의 재범위험성을 예측하는 인공지능 알고리즘이 개발되어 형사절차에 실제로 활용되는 단계에 이르렀다. 이러한 알고리즘의 활용은 수사기관이나 재판기관의 자의적 판단을 방지하고 판단의 객관성과 합리성을 제고하는 방법이 될 수도 있는 반면, 데이터의 오류로 인해 왜곡된 판단이 도출될 가능성, 기존의 편견을 강화ㆍ재생 산할 위험성, 예측의 부정확성, 알고리즘의 작동 방식 비공개로 인한 검증 불가능성, 법관을 비롯한 의사결정주체들이 알고리즘의 판단결과에 지나친 영향을 받을 위험성 등이 문제점으로 지적 된다. 형사절차에서 인공지능 알고리즘의 활용가능성과 한계는 구체적인 형사절차상의 국면에 따라 달리 판단되어야 한다. 인공지능 알고리즘의 예측 결과는 경찰관이 불심검문 대상자 여부를 판단할 때나 법원이 피의자의 인신구속 여부를 판단할 때 근거자료로 활용될 수 있을 것이나, 개별적ㆍ 구체적인 정황 없이 오로지 알고리즘의 예측결과만을 근거로 불심검문을 시행하거나 인신구속을 할 수는 없다고 보아야 한다. 공판절차에서의 유무죄 판단의 경우에는 엄격한 증명의 법리가 적용되는데, 현재 개발ㆍ활용되고 있는 인공지능 예측기술들은 아직 그 이론적 근거나 정확성이 충분히 검증되어 일반적으로 승인되는 수준에 이르렀다고 할 수 없으므로 증거능력을 인정할 수 없다고 보아야 한다. 엄격한 증명의 법리가 적용되지 않는 양형 판단의 경우에는 인공지능 알고리즘에 의한 재범위험성 평가결과를 판단의 근거자 료로 참작할 수 있을 것이나, 이 경우 헌법상 피고인에게 보장되는 법관에 의한 공정한 재판을 받을 권리에 의한 제한을 받는다. 따라서 알고리즘에 의한 자료를 제출할 때는 반드시 알고리즘의 정확성에 대한 검증결과를 함께 제출하게 하여야 하고, 알고리즘의 작동원리를 공개하고 이를 법관과 당사자들이 이해할 수 있는 방식으로 설명하도록 하여야 하며, 알고리즘의 판단결과에 대해 피고인이 실질적으로 반박할 수 있는 기회를 보장하여야 할 것이다. 입법론으로는 인공지능 알고리즘의 형사절차에서의 활용에 관한 기준과 한계를 법령으로 명확하게 정하는 것이 바람직하다.

Anticipation of welding deformation with finite element method is a very interested topic in the industries, adequate heat source model is essential for concluding reasonable results. This study is related to estimate the parameters of Goldak heat source model, and global optimization algorithm is applied to this research. The heat affected zone (HAZ) boundary line of bead on plate (BOP) welding is used as the target, parameters of heat sources are used as the variables. Adaptive simulated annealing is applied and the optimal result is obtained out of 1,000 candidates. The convergence of finite element method and the global optimization is meaningful for estimation of welding deformation, which could enhance to reduce the resources and time for experiments.

본 논문은 유한요소법과 유전알고리즘을 연동하여 지진하중을 받는 구조물의 강성저하(손상) 및 보강 후 효과를 추정하는 방법을 다루었다. 본 연구의 독창성은 지진하중을 적용하였고, 그 응답으로부터 구조물의 미지 변수를 추정한다는 점이다. 본 연구에서 제안한 방법은 지진하중으로부터 손상된 부위를 추정할 뿐 아니라, 그 위치와 정도를 규명할 수 있다. 제안한 방법을 검증하기 위하여 El Centro 및 포항 지진하중을 적용하여 저층 뼈대구조물와 트러스 교량을 대상으로 알고리즘을 실행하였다. 수치해석 예제는 제안한 방법이 수치해석적인 효율성 뿐 아니라 지진으로부터의 심각한 피해를 예방하는 데 적용할 수 있음을 보여주었다.

In this study, a controlled atmosphere (CA) storage system is proposed as a storage method for prolonging processing period. Persimmon was placed in CA storage at 0.5oC±0.5 for 92 d. The qualities of the stored persimmons were compared to determine the possibility of extending shelf life. ‘Sangjudungsi,’ which was harvested on October 26, 2017, was applied to the persimmons. In order to compare differences according to size, the persimmons were classified into Size No. 2 (170 g) and Size No. 3 (145 g). In the result, the yellowness of CA-stored persimmons was 26.3% lower than that of cold-stored ones, confirming that after-ripening was delayed. The firmness of CAstored and Size No. 3 persimmon was higher than that of cold-stored and Size No. 2 persimmon. Tannin decreased significantly in cold storage, but it tended to increase in CA storage. The sugar content of Size No. 3 was lower than that of Size No. 2, but there was no difference in tendencies according to the storage method. Weight loss in CA storage was lower than that in cold storage. A comparison of color difference, firmness, sugar content, tannin, and weight loss ratio showed that CA storage was more effective in improving shelf life than cold storage.

선박이 접안할 때 발생하는 접안에너지에 가장 영향력이 큰 요소는 접안속도이며, 과도한 경우 사고로 이어질 수 있다. 접안속도의 결정에 영향을 미치는 요소는 다양하지만 기존 연구에서는 일반적으로 선박 크기에 제한하여 분석하였다. 따라서 본 연구에서는 다양한 선박 접안속도의 영향요소를 반영하여 분석하고 그에 따른 중요도를 도출하고자 한다. 분석에 활용한 데이터는 국내 한 탱커부두의 선박 접안속도를 실측한 것을 바탕으로 하였다. 수집된 데이터를 활용하여 머신러닝 분류 알고리즘인 의사결정나무(Decision Tree), 랜덤포 레스트(Random Forest), 로지스틱회귀(Logistic Regression), 퍼셉트론(Perceptron)을 비교분석하였다. 알고리즘 평가 방법으로는 혼동 행렬에 따른 모델성능 평가지표를 사용하였다. 분석 결과, 가장 성능이 좋은 알고리즘으로는 퍼셉트론이 채택되었으며 그에 따른 접안속도 영향 요인의 중요도는 선박 크기(DWT), 부두 위치(Jetty No.), 재화상태(State) 순으로 나타났다. 이에 따라 선박 접안 시, 선박의 크기를 비롯하여 부두 위치, 재화 상태 등 다양한 요인을 고려하여 접안속도를 설계하여야 한다.

The structural design of arch roofs or bridges requires the analysis of their unstable behaviors depending on certain parameters defined in the arch shape. Their maintenance should estimate the parameters from observed data. However, since the critical parameters exist in the equilibrium paths of the arch, and a small change in such the parameters causes a significant change in their behaviors. Thus, estimation to find the critical ones should be carried out using a global search algorithm. In this paper we study the parameter estimation for a shallow arch by a quantum-inspired evolution algorithm. A cost functional to estimate the system parameters included in the arch consists of the difference between the observed signal and the estimated signal of the arch system. The design variables are shape, external load and damping constant in the arch system. We provide theoretical and numerical examples for estimation of the parameters from both contaminated data and pure data.

PURPOSES: The intensiveness of highway management has increased owing to the growth in the number of vehicles and the rapid climate change. The disadvantages produced by these factors can affect management time and cost. Serious traffic accidents and traffic jam may be experienced when snow fall accumulates on highway surfaces and the friction between tires and pavements is lower than that in the general state, in a non-management condition. Such conditions need intensive management. In this regard, one of the spread methods used for the melting material is pre-wetted salt (PWS), which is the frequently used method in South Korea. In the PWS method, the solid material with CaCl2 is mixed with water in 30% concentration and then finally mixed with NaCl before application to pavements. The chloride-type melting material not only is cheaper, but also has a high melting property than the others. It can shorten the pavement or structure life by deterioration and corrosion. This melting material can affect the flora near the highways; hence, an eco-friendly de-icing agent must be utilized considering the environmental effect. METHODS : The Kalman filter algorithm (KFA) was utilized herein to develop optimization models using the performed test data. The KFA, which was developed from recursive filter algorithms, such as the low- and high-pass filters, applies a weighting filter to the Kalman filter. The algorithm has the property of utilizing the filter and updated estimations. In this regard, melting tests were performed for the real applicative utilization of de-icing agents. The KFA was also applied to reduce the error rates and optimize the relationships between the test data and the predictions. RESULTS: Comparing the measurements performed, the error was reduced by 1.69 g when the KFA was applied. Moreover, the error can be optimized to approximately 91.4% compared to the test errors. The prediction data had over 85% tendency in the test measurement, showing that the KFA application can reduce the error and increase the tendency. By comparison, the agent with CaCl2 showed the best ice melting performance within 10 min without surface temperature. However, the PWS with a 25% concentration indicated the best water melting performance from start to end of the test time, implying that this is a powerful agent in terms of performance. CONCLUSIONS : The melting test is an artificial test method; therefore, it can generate a huge error from the test. The error and the tendency can be controlled by tracking the measurement error and the white noise matrix using the KFA. A further research will be performed to track the measurement error and the white noise matrix. Other optimization methods will also be applied to reduce the experimental error.

Process mining is an analytical technique aimed at obtaining useful information about a process by extracting a process model from events log. However, most existing process models are deterministic because they do not include stochastic elements such as the occurrence probabilities or execution times of activities. Therefore, available information is limited, resulting in the limitations on analyzing and understanding the process. Furthermore, it is also important to develop an efficient methodology to discover the process model. Although genetic process mining algorithm is one of the methods that can handle data with noises, it has a limitation of large computation time when it is applied to data with large capacity. To resolve these issues, in this paper, we define a stochastic process tree and propose a tabu search-genetic process mining (TS-GPM) algorithm for a stochastic process tree. Specifically, we define a two-dimensional array as a chromosome to represent a stochastic process tree, fitness function, a procedure for generating stochastic process tree and a model trace as a string of activities generated from the process tree. Furthermore, by storing and comparing model traces with low fitness values in the tabu list, we can prevent duplicated searches for process trees with low fitness value being performed. In order to verify the performance of the proposed algorithm, we performed a numerical experiment by using two kinds of event log data used in the previous research. The results showed that the suggested TS-GPM algorithm outperformed the GPM algorithm in terms of fitness and computation time.

The Dynamic Vehicle Routing Problem (DVRP) involves a combinatorial optimization problem where new customer demands become known over time, and old routes must be reconfigured to generate new routes while executing the current solution. We consider the high level of dynamism problem. An application of highly dynamic DVRP is the ambulance service where a patient contacts the service center, followed by an evaluation of case severity, and a visit by a practitioner/ ambulance is scheduled accordingly. This paper considers a variant of the DVRP and proposes a decentralized algorithm in which collaborators (Depot and Vehicle), both have only partial information about the entire system. The DVRP is modeled as a periodic re optimization of VRP using the proposed decentralized algorithm where collaborators exchange local information to achieve the best global objective for the current state of the system. We assume the existence of a dispatcher e.g., headquarter of the company who can communicate to vehicles in order to gather information and assigns the new visits to them. The effectiveness of the proposed decentralized coordination algorithm is further evaluated using benchmark data given in literature. The results show that the proposed method performed better than the compared algorithms which utilize the centralized coordination in 12 out of 21 benchmark problems.

The performance of ground-based optical structures is highly sensitive to external environments, such as airflow in open space. In this paper, initial aerodynamic data due to ambient air flow were analyzed in optical models designed through knowledge-based design algorithm, and dynamic data acting on optical structures in turbulent flow with velocity of 50m/s were analyzed to present the initial shape design conditions of the structures. The simulation results showed that the maximum pressure, minimum pressure, and maximum differential pressure acting on the mirror are directly proportional to the sweep angle.

신재생 에너지 자원중 풍력발전은 비약적인 기술 발전과 시장 규모가 급속하게 성장하고 있다. 최근 육상풍력발전단지의 공간적 한계, 환경 문제 등으로 인하여 설치 공간이 해상으로 이동되었고, 더욱 풍부한 풍황 조건을 가진 깊은 수심에 설치되는 부유식 해상 풍력단지의 개발이 활발하게 진행되고 있다. 해상교통관점에서 해상풍력단지의 최적위치 선정은 선박과 풍력기들의 간섭을 최소화 하고 사고 확률이 적은 곳이며, 선박 밀집도가 낮은 해역이 최적위치로 선정된다. 본 연구에서는 유전 알고리즘 기반의 계절별 1주일 기간 선박자동식별장치 데이터를 유전자 및 염색체로 구성하였다. 80개의 유전자로 구성하고 유전 알고리즘의 적합도 평가를 거쳐 부유식 해상 풍력단지의 계절별 최적위치를 선정하였다. 더 나아가 계절별 최적위치 점수를 합산하여 최종 최적위치를 선정하였다. 분석 해역에서 최적위치는 11개로 나타났으며, 해상교통관점에서 유전 알고리즘을 통한 최적위치 선정이 적용 가능함을 확인하였다.

The concept of carsharing involves sharing a small number of reserved cars to be used individually by a larger number of people as required. This study examines the operating parameters of one-way carsharing systems in order to determine the appropriate operating conditions that minimizes the lost sales rate. Five operating parameters are tested in this study: the number of stations, the average number of vehicles per station, the rate of one-way trip, the average number of staffs per station, and the relocation policy. The performance of round-trip carsharing systems is also compared to that of one-way carsharing systems. A simulation model is developed and simulations are performed to determine the appropriate combination of operating parameter and levels. The simulation results show that the average number of vehicles per station is the most critical parameter. Other key findings obtained from this research are as follows. First, applying the appropriate relocation policy to one-way carsharing systems can allow more customers to rent vehicles than the traditional round-trip carsharing systems. Second, the appropriate relocation policy should be selected based on the average number of vehicles per station in order to minimize the lost sales rate. Third, the number of stations does not affect the lost sales rate. This study findings will provide tools to understand impact of the carsharing system parameters on the efficiency of the carsharing operations.