• Proceedings of the KIEE Conference
• /
• 2005.07d
• /
• pp.2933-2935
• /
• 2005

The IRPO(Intensive Randomized Policy Optimizer) algorithm is a recently developed tool in the area of reinforcement leaming. And it has been shown to be very successful in several application problems. To compare with a general RL method, IRPO has some difference in that policy utilizes the entire history of agent -environment interaction. The policy is derived from the history directly, not through any kind of a model of the environment. In this paper, we consider a robot-control problem utilizing a IRPO algorithm. We also developed a MATLAH-based animation program, by which the effectiveness of the training algorithms were observed.

• Journal of Fisheries and Marine Sciences Education
• /
• v.27 no.4
• /
• pp.1160-1171
• /
• 2015

The purpose of this study is to design and develop a computer based learning and test system, which supports not only testing learner's ability but also learning contents with giving feedback and hint. In order to design and develop a computer based learning and test system, Visual Basic dot Net software is used. The system works in three stages: sequential problem solving stage, randomized problem solving stage, and the challenge stage of pass/fail. The results of this study are as follows: (a) We propose the context of design for the computer based learning and test system. (b) We design and develop items display function with sequential and random algorithm in this system. (c) We design and develop pass/fail function by applying SPRT(Sequential Probability Ratio Testing) algorithm in the computer based learning and test system.

• Journal of Korean Institute of Industrial Engineers
• /
• v.29 no.3
• /
• pp.222-229
• /
• 2003

In this paper, we address a layout design problem, PTN[2], for determining an appropriate 2-class-based dedicated storage layout in a class of unit load storage systems. Our strong conjecture is that PTNI2] is NP-hard. Restricting PTN[2], we provide three solvable cases of PTN[2] in which an optimal solution to the solvable cases is one of the partitions based on the PAI(product activity index)-nonincreasing ordering. However, we show with a counterexample that a solution based on the PAI-non increasing ordering does not always give an optimal solution to PTN[2]. Utilizing the derived properties, we construct an effective heuristic algorithm for solving PTN[2] based on a PAI-non increasing ordering with performance ratio bound. Our algorithm with O($n^2$) is effective in the sense that it guarantees a better class-based storage layout than a randomized storage layout in terms of the expected single command travel time.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.5C
• /
• pp.415-421
• /
• 2005

Passive RFID tag does not have its own power, so it has very poor computation abilities and it can deliver signals in very short range. From the facts, most RFID Tag security schemes assumed that the backward channel from tags to a reader is safe from eavesdropping. However, eavesdroppers near a tag can overhear message from a tag illegally. In this paper, we propose a method to protect the backward channel from eavesdropping by illegal readers. The proposed scheme can overcome the problems of conventional schemes such as randomized tree walking, which have been proposed to secure tag information in tree-walking algorithm as an anti-collision scheme for RFID tags. We showed the efficiency of our proposed method by using an analytical model, and it is also shown that the proposed method can provide the probability of eavesdropping in some standardized RFID tag system such as EPCglobal, ISO, uCode near to '0'.

• Journal of Haehwa Medicine
• /
• v.24 no.1
• /
• pp.15-24
• /
• 2015

Objectives : This study was carried out to assess the risk of bias of clinical trials on acne treatment with herbal medicine that have been published in Korea. Methods : 7 electronic databases in Korea were searched for clinical trials on acne treatment. Two independent reviewers selected clinical trials on herbal medicine treatment for acne. Selected studies are categorized according to DAMI(Study Design Algorithm for Medical literature of Intervention). RCTs are assessed according to Cochrane RoB(Risk of Bias), non-randomized studies(Before-after studies) are assessed according to RoBANS(Risk of Bias Assessment tool for Non-randomized Study). Results : After selection process, 25 articles are left. Among 25 articles, 3 RCTs and 4 before-after studies are finally included. In RCTs, the proportion of 'unclear' is high in criteria of 'random sequence generation', 'allocation concealment', and 'blinding'. In before-after studies, 'high' is high in criteria of 'blinding for outcome assessment' and 'incomplete outcome data'. Conclusions : Considering the above results of the assessment, it is necessary to conduct more well designed clinical trials on acne treatment with herbal medicine.

• Journal of KIISE:Software and Applications
• /
• v.28 no.9
• /
• pp.662-668
• /
• 2001

Since the problem of disproving a tautology is as hard as the problem of proving it, no polynomial time algorithm for falsification(or testing invalidity) is feasible. Previous algorithms are mostly based on either divide-and-conquer or graph representation. Most of them demonstrated satisfactory results on a variety of input under certain constraints. However, they have experienced difficulties dealing with big input. We propose a new falsification algorithm using a Merge Rule to produce a counterexample by constructing a minterm which is not satisfied by an input expression in DNF(Disjunctive Normal Form). We also show that the algorithm is consistent and sound. The algorithm is based on a greedy method which would seek to maximize the number or terms falsified by the assignment made at each step of the falsification process. Empirical results show practical performance on big input to falsify randomized nontautological problem instances, consuming O(nm$^2$) time, where n is the number of variables and m is number of terms.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.4B
• /
• pp.338-348
• /
• 2009

The Wireless Sensor Network (WSN) has recently attracted considerable attention due to the low price and ease to deploy it. In particular, in a hostile or harsh regions where sensors cannot be deployed manually, WSNs can be established just by dropping sensors from the air. In this case, however, most likely sensors are not placed at optimal positions, although the location of sensors does have a drastic impact on the WSN performance. Moreover, randomized deployment algorithm can leave holes in terms of coverage in the sensing area. This paper proposes a sensor relocation scheme where mobile sensors move to patch up the holes by appropriate coverage. Simulation results show that the proposed algorithm outperforms prior existing schemes in terms of coverage and lifespan of WSNs.

• Nuclear Engineering and Technology
• /
• v.56 no.2
• /
• pp.558-567
• /
• 2024

Diagnostic support systems are being researched to assist operators in identifying and responding to abnormal events in a nuclear power plant. Most studies to date have considered single abnormal events only, for which it is relatively straightforward to obtain data to train the deep learning model of the diagnostic support system. However, cases in which multiple abnormal events occur must also be considered, for which obtaining training data becomes difficult due to the large number of combinations of possible abnormal events. This study proposes an approach to maintain diagnostic performance for multiple abnormal events by training a deep learning model with data on single abnormal events only. The proposed approach is applied to an existing algorithm that can perform feature selection and multi-label classification. We choose an extremely randomized trees classifier to select dedicated monitoring parameters for target abnormal events. In diagnosing each event occurrence independently, two-channel convolutional neural networks are employed as sub-models. The algorithm was tested in a case study with various scenarios, including single and multiple abnormal events. Results demonstrated that the proposed approach maintained diagnostic performance for 15 single abnormal events and significantly improved performance for 105 multiple abnormal events compared to the base model.

• Clinical Endoscopy
• /
• v.57 no.3
• /
• pp.283-292
• /
• 2024

We developed three e-learning systems for endoscopists to acquire the necessary skills to improve the diagnosis of early gastric cancer (EGC) and demonstrated their usefulness using randomized controlled trials. The subjects of the three e-learning systems were "detection", "characterization", and "preoperative assessment". The contents of each e-learning system included "technique", "knowledge", and "obtaining experience". All e-learning systems proved useful for endoscopists to learn how to diagnose EGC. Lecture videos describing "the technique" and "the knowledge" can be beneficial. In addition, repeating 100 self-study cases allows learners to gain "experience" and improve their diagnostic skills further. Web-based e-learning systems have more advantages than other teaching methods because the number of participants is unlimited. Histopathological diagnosis is the gold standard for the diagnosis of gastric cancer. Therefore, we developed a comprehensive diagnostic algorithm to standardize the histopathological diagnosis of gastric cancer. Once we have successfully shown that this algorithm is helpful for the accurate histopathological diagnosis of cancer, we will complete a series of e-learning systems designed to assess EGC accurately.

• The KIPS Transactions:PartC
• /
• v.17C no.3
• /
• pp.291-298
• /
• 2010

In wireless sensor networks, energy is the most important consideration because the lifetime of the sensor node is limited by battery. The clustering is the one of methods used to manage network energy consumption efficiently and LEACH(Low-Energy Adaptive Clustering Hierarchy) is one of the most famous clustering algorithms. LEACH utilizes randomized rotation of cluster-head to evenly distribute the energy load among the sensor nodes in the network. The random selection method of cluster-head does not guarantee the number of cluster-heads produced in each round to be equal to expected optimal value. And, the cluster head in a high-density cluster has an overload condition. In this paper, we proposed both a token based cluster-head selection algorithm for guarantee the number of cluster-heads and a cluster selection algorithm for uniform-density cluster. Through simulation, it is shown that the proposed algorithm improve the network lifetime about 9.3% better than LEACH.