• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.1
• /
• pp.171-176
• /
• 2020

Reinforcement learning is a well-known method for training an artificial software agent for a video game. Even though many reinforcement learning algorithms have been proposed, their performance was varies depending on an application area. This paper compares the performance of the algorithms when we train our reinforcement learning agent for a 2D racing game. We defined performance metrics to analyze the results and plotted them into various graphs. As a result, we found ACER (Actor Critic with Experience Replay) achieved the best rewards than other algorithms. There was 157% gap between ACER and the worst algorithm.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.11 no.5
• /
• pp.459-464
• /
• 2016

The mobile communication service is widely used and the repeaters is rapidly increasing because of the extending service areas. we propose a new Signed CMA(: Constant Modulus Algorithm) algorithm for ICS(: Interference Cancellation System) as a larger service area of mobile communication service. The proposed Signed CMA algorithm improved performances by adjusting step size values. We showed that the proposed algorithm could improve the Mean Square Error(: MSE) performance of Signed CMA. MATLAB(: Matrix Laboratory) is employed to analyze the proposed algorithm and to compare it with the experimental results. At the convergence of 1000 iteration state, the MSE(: Mean Square Error) performance of the proposed Signed CMA algorithm with step size of 0.0065 is about 5 dB better than the conventional CMA algorithm.

• Journal of Korea Multimedia Society
• /
• v.13 no.5
• /
• pp.763-770
• /
• 2010

This paper proposes an algorithm to enhance the performance of the spatial partitioning method for a flocking behavior. One of the characteristics in a flocking behavior is that two agents may share many common neighbors if they are spatially close to each other. This paper improves the spatial partitioning method by applying this characteristic. While the conventional spatial partitioning method computes the k-nearest neighbors of an agent one by one, the proposed method computes simultaneously the k-nearest neighbors of agents if they are spatially close to each other. The proposed algorithm was implemented and its performance was experimentally compared with the original spatial partitioning method. The results of the comparison showed that the proposed algorithm outperformed the original method by about 33% in average.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.6
• /
• pp.89-97
• /
• 2016

This paper addresses the problem of the construction planning of artillery positions, for which we present an optimization model and propose a heuristic algorithm to solve problems of practical size. The artillery position construction plan includes the assignment of engineers to support the artillery and the schedule of the support team construction sequence. Currently, in the army, managers construct the plan based on their experience. We formulate the problem as a mixed integer program and present a heuristic that utilizes the decomposition of the mixed integer model. We tested the efficacy of the proposed algorithm by conducting computational experiments on both small-size test problems and large-size practical problems. The average optimality gap in the small-size test problem was 6.44% in our experiments. Also, the average computation time to solve the large-size practical problems consisting of more than 200 artillery positions was 79.8 seconds on a personal computer. The result of our computational experiments shows that the proposed approach is a viable option to consider for practical use.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.6
• /
• pp.2263-2268
• /
• 2010

In this paper, the hybrid compensation algorithm($A_{HB}$) for localization using the Compensation Algorithm distance($CA_d$) and the Algorithm of Equivalent Distance Rate(AEDR) in SDS-TWR(Symmetric Double-Sided Two-Way Ranging) is suggested and the performance of the proposed algorithm is analyzed by practical experimentations. From experimentations, it is confirmed that the errors are reduced in 28 coordinates of total 32 coordinates in the experimental region and the errors are reduced about above 70% in the assigned 3 type error level ranges by $A_{HB}$. Also, it is analyzed that the average localization error is reduced from 2.67m to 1.19m as 55.4% in total 32 coordinates by $A_{HB}$ and the error compensation capability of $A_{HB}$ is very excellent as above 90%. From above results, we have seen that the error reduction ratio and error compensation capability of $A_{HB}$ is more excellent than each $CA_d$ or AEDR.

• Journal of Broadcast Engineering
• /
• v.22 no.3
• /
• pp.381-396
• /
• 2017

In this paper, the perceptual rate control algorithm is studied for HEVC (High Efficiency Video Coding) encoder with bit allocation based on perceived visual quality. This paper proposes perceptual rate control algorithm which could consider perceived quality for HEVC encoding method. The proposed rate control algorithm employs adaptive bit allocation for frame and CTU level using the perceived visual importance of each CTU. For performance evaluation of the proposed algorithm, the proposed algorithm was implemented on HM 16.9 and tested for sequences in Class B under the CTC (Common Test Condition) RA (Random Access) case. Experimental results show that the proposed method reduces the bitrate of 3.12%, and improves BD-PSNR of 0.08dB and bitrate accuracy of 0.07% on average. And also, we achieved MOS improvement of 0.16 with the proposed method, compared with the conventional method based on DSCQS (Double Stimulus Continuous Quality Scale).

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.12
• /
• pp.2769-2774
• /
• 2013

The key of USN(Ubiquitous Sensor Network) technology is low power wireless communication technology and proper resource allocation technology for efficient routing. The distinguished resource allocation method is needed for efficient routing in sensor network. To solve this problems, we propose an algorithm that can be adopted in USN with making up for weak points of PQ and WRR in this paper. The proposed algorithm produces the control discipline by the fuzzy theory to dynamically assign the weight of WRR scheduler with checking the Queue status of each class in sensor network. From simulation results, the proposed algorithm improves the packet loss rate of the EF class traffic to 6.5% by comparison with WRR scheduling method and that of the AF4 class traffic to 45% by comparison with PQ scheduling method.

• Journal of Advanced Navigation Technology
• /
• v.18 no.5
• /
• pp.461-467
• /
• 2014

Active queue management (AQM) algorithms run on routers and detect incipient congestion by typically monitoring the instantaneous or average queue size. When the average queue size exceeds a certain threshold, AQM algorithms infer congestion on the link and notify the end systems to back off by proactively dropping some of the packets arriving at a router or marking the packets to reduce transmission rate at the sender. Among the existing AQM algorithms, random early detection (RED) is well known as the representative queue-based management scheme by randomizing packet dropping. To reduce the number of timeouts in TCP and queuing delay, maintain high link utilization, and remove bursty traffic biases, the RED considers an average queue size as a degree of congestions. However, RED do not well in the specified networks conditions due to the fixed parameters($P_{max}$ and $TH_{min}$) of RED. This paper addresses a extended RED to be adapted in various networks conditions. By sensing network state, $P_{max}$ and $TH_{min}$ can be automatically changed to proper value and then RED do well in various networks conditions.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.1
• /
• pp.49-54
• /
• 2016

In this paper, we propose a vocal separation method using weighted ${\beta}$-order minimum mean wquare error estimation (WbE) based on kernel back-fitting algorithm. In spoken speech enhancement, it is well-known that the WbE outperforms the existing Bayesian estimators such as the minimum mean square error (MMSE) of the short-time spectral amplitude (STSA) and the MMSE of the logarithm of the STSA (LSA), in terms of both objective and subjective measures. In the proposed method, WbE is applied to a basic iterative kernel back-fitting algorithm for improving the vocal separation performance from monaural music signal. The experimental results show that the proposed method achieves better separation performance than other existing methods.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2743-2745
• /
• 2002

역전파 신경망과 데이터분포 특징을 고려한 새로운 알고리즘을 개발하였으며, 이를 플라즈마 데이터의 분류에 응용하였다. 데이터 분포는 통계적인 평균치와 표준편차를 이용하여 특징지었으며, 바이어스인자를 이용하여 9 종류의 데이터를 발생하였다. 각 데이터에 대하여 은닉층의 뉴런수를 변화시키며, 바이어스와 뉴런수에 따른 모델성능을 평균학습시간 (ATT), 평균예측정확도 (APA), 최적예측정확도 (BPA), 그리고 분류정확도 (CA) 측면에서 세분하여 분석하였다. ATT와 APA에 대해서는 최적화된 학습인자와 데이터 분류인자가 일치하였고, BPA와 CA는 일치하지 않았다. 두 인자간의 상호작용을 동시에 최적화함으로써 완전 분류를 달성하였다.