• 한국경영과학회지
• /
• 제21권1호
• /
• pp.101-114
• /
• 1996

This study suggests an accelerated genetic algorithm for the vehicle routing problem (AGAVRP). This algorithm treats both the single-visit and the multiple-visit models. AGAVRP is accelerated by the OR techniques at the various stages of the algorithm. In order to improve the convergence of AGAVRP, a robust set of parameters is determined by the experimental design approach. The relative performance of AGAVRP is comparable to the other known algorithms. The advantage of the proposed algorithm is flexibility and better convergence.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제13권4호
• /
• pp.284-290
• /
• 2013

Feature detection is very important to image processing area. In this paper we compare and analyze some characteristics of image processing algorithms for corner and blob feature detection. We also analyze the simulation results through image matching process. We show that how these algorithms work and how fast they execute. The simulation results are shown for helping us to select an algorithm or several algorithms extracting corner and blob feature.

• Journal of information and communication convergence engineering
• /
• 제20권2호
• /
• pp.96-102
• /
• 2022

We aim to build predictive models for Airbnb's prices using a GPU-accelerated RAPIDS in a big data cluster. The Airbnb Listings datasets are used for the predictive analysis. Several machine-learning algorithms have been adopted to build models that predict the price of Airbnb listings. We compare the results of traditional and big data approaches to machine learning for price prediction and discuss the performance of the models. We built big data models using Databricks Spark Cluster, a distributed parallel computing system. Furthermore, we implemented models using multiple GPUs using RAPIDS in the spark cluster. The model was developed using the XGBoost algorithm, whereas other models were developed using traditional central processing unit (CPU)-based algorithms. This study compared all models in terms of accuracy metrics and computing time. We observed that the XGBoost model with RAPIDS using GPUs had the highest accuracy and computing time.

• 터널과지하공간
• /
• 제6권4호
• /
• pp.316-325
• /
• 1996

A new accelerated neural network adopting modified sigmoid function was developed and applied to estimate engineering properties of rock from insufficient geological data. Developed network was tested on the well-known XOR and character recognition problems to verify the validity of the algorithms. Both learning speed and recognition rate were improved. Test learn on the Lee and Sterling's problems showed that learning time was reduced from tens of hours to a few minutes, while the output pattern was almost the same as other studies. Application to the various case studies showed exact coincidence with original data or measured results.

• Nonlinear Functional Analysis and Applications
• /
• 제27권3호
• /
• pp.553-568
• /
• 2022

In this paper, we introduce and study two different iterative hybrid projection algorithms for solving a fixed point problem of nonexpansive mappings. The first algorithm is generated by the combination of the inertial method and the hybrid projection method. On the other hand, the second algorithm is constructed by the convex combination of three updated vectors and the hybrid projection method. The strong convergence of the two proposed algorithms are proved under very mild assumptions on the scalar control. For illustrating the advantages of these two newly invented algorithms, we created some numerical results to compare various numerical performances of our algorithms with the algorithm proposed by Dong and Lu [11].

• 한국방사선학회논문지
• /
• 제17권5호
• /
• pp.633-640
• /
• 2023

최근 근접 치료에서 방사선 차폐막을 사용하여 선량 분포를 변조하여 선량을 전달하는 정적 및 동적 변조 근접 치료 방법이 개발됨에 따라 새로운 방향성 빔 세기 변조 근접 치료에 적합한 역방향 치료 계획 및 치료 계획 최적화 알고리즘에서 선량 계산에 필요한 파라미터 및 데이터의 양이 증가하고 있다. 세기 변조 근접 치료는 방사선의 정확한 선량 전달이 가능하지만, 파라미터와 데이터의 양이 증가하기 때문에 선량 계산에 필요한 경과 시간이 증가한다. 본 연구에서는 선량 계산 경과 시간의 증가를 줄이기 위해 그래픽 카드 기반의 CUDA 가속 선량 계산 알고리즘을 구축하였다. 계산 과정의 가속화 방법은 관심 체적의 시스템 행렬 계산 및 선량 계산의 병렬화를 이용하여 진행하였다. 개발된 알고리즘은 모두 인텔(3.7GHz, 6코어) CPU와 단일 NVIDIA GTX 1080ti 그래픽 카드가 장착된 동일한 컴퓨팅 환경에서 수행하였으며, 선량 계산 시간은 디스크에서 데이터를 불러오고 전처리를 위한 작업 등의 추가 적으로 필요한 시간은 제외하고 선량 계산 시간만 측정하여 평가하였다. 그 결과 가속화된 알고리즘은 CPU로만 계산할 때보다 선량 계산 시간이 약 30배 단축된 것으로 나타났다. 가속화된 선량 계산 알고리즘은 적응방사선치료와 같이 매일 변화되는 어플리케이터의 움직임을 고려하여 새로운 치료 계획을 수립해야 하는 경우나 동적 변조 근접 치료와 같이 선량 계산에 변화되는 파라미터를 고려해야 하는 경우 치료 계획 수립 속도를 높일 수 있을 것으로 판단된다.

• International Journal of Aeronautical and Space Sciences
• /
• 제3권1호
• /
• pp.50-60
• /
• 2002

A new co-evolutionary algorithm, of which the convergence speed is accelerated by neural networks, is proposed and verified in this paper. To reduce computational load required for co-evolutionary optimization processes, the cost function and constraint information is stored in the neural networks, and the extra offspring group, whose cost is computed by the neural networks, is generated. It increases the offspring population size without overloading computational effort; therefore, the convergence speed is accelerated. The proposed algorithm is applied to attitude control design of flexible satellites, and it is verified by computer simulations and experiments using a torque-free air bearing system.

• 대한전자공학회논문지SP
• /
• 제49권2호
• /
• pp.28-34
• /
• 2012

특징 추출 알고리즘은 영상 내에서 중요한 특징을 추출하기 위해 실시간 영상 처리 응용 분야에서 활용된다. 특히, 특징 추출 알고리즘은 추적 및 식별의 목적으로 다양한 영상처리 알고리즘에 특징 정보를 제공하기 위해서 활용되며, 주로 영상처리 전처리 단계에서 구현되고 있다. 광범위한 응용 분야에 이용되는 특징 추출 알고리즘의 처리 속도를 높인다면 혼합되어 사용될 다른 알고리즘 처리 소요 시간의 여유를 확보 할 수 있을 뿐만 아니라, 특징 추출 알고리즘이 적용된 영상 처리 응용 분야의 실시간 요건을 만족시키기 용이하기 때문에 중요하다. 본 논문에서는 특징 추출 기법을 고속으로 처리하기 위해 FPGA 기반의 하드웨어 가속기를 제안한다. 하드웨어 가속기 구현에 사용된 E. Rosten의 Feature from Accelerated Segment Test 알고리즘과 디지털 로직으로 구현한 하드웨어 가속기의 구조와 동작 절차에 대해 기술하였다. 설계한 하드웨어 가속기는 ModelSim을 이용해 동작 및 성능을 검증하였고, Xilinx Vertex IV FPGA 기반으로 로직을 합성해 구현 비용을 계산하였다. 제안한 하드웨어 가속기를 구현하기 위해 2,217개의 Flip Flop, 5,034개의 LUT, 2,833개의 Slice, 그리고 18개의 Block RAM을 사용하였으며, $640{\times}480$ 크기의 영상으로부터 954개의 특징을 추출하는데 3.06 ms의 시간이 소요되어 기존의 결과보다 구현 비용 면에서의 우월함이 확인되었다.

• 로봇학회논문지
• /
• 제17권2호
• /
• pp.133-141
• /
• 2022

This paper introduces model compression algorithms which make a deep neural network smaller and faster for embedded systems. The model compression algorithms can be largely categorized into pruning, quantization and knowledge distillation. In this study, gradual pruning, quantization aware training, and knowledge distillation which learns the activation boundary in the hidden layer of the teacher neural network are integrated. As a large deep neural network is compressed and accelerated by these algorithms, embedded computing boards can run the deep neural network much faster with less memory usage while preserving the reasonable accuracy. To evaluate the performance of the compressed neural networks, we evaluate the size, latency and accuracy of the deep neural network, DenseNet201, for image classification with CIFAR-10 dataset on the NVIDIA Jetson Xavier.

• Nonlinear Functional Analysis and Applications
• /
• 제29권2호
• /
• pp.307-332
• /
• 2024

Two inertial extragradient-type algorithms are introduced for solving convex pseudomonotone variational inequalities with fixed point problems, where the associated mapping for the fixed point is a 𝜌-demicontractive mapping. The algorithm employs variable step sizes that are updated at each iteration, based on certain previous iterates. One notable advantage of these algorithms is their ability to operate without prior knowledge of Lipschitz-type constants and without necessitating any line search procedures. The iterative sequence constructed demonstrates strong convergence to the common solution of the variational inequality and fixed point problem under standard assumptions. In-depth numerical applications are conducted to illustrate theoretical findings and to compare the proposed algorithms with existing approaches.