• 한국전산구조공학회:학술대회논문집
• /
• 한국전산구조공학회 1999년도 봄 학술발표회 논문집
• /
• pp.50-57
• /
• 1999

Structural analysis of large-scale structures involving large amount of computational load and data storage requires high-performance computing resources. We have previously developed PC-level distributed structural analysis algorithms based on substructuring technique where each personal computer assigned to a slave node has been involved in the computations for single substructures. Recently, it has been proved by the authors that the performance of distributed structural analysis algorithm can be further enhanced by changing substructuring schemes. Therefore a new distributed structural analysis algorithm with one PC to multiple substructures scheme is presented in this paper. The algorithm is implemented on the network of multiple personal computers and applied to structural analysis of two dimensional frame structures.

• Journal of information and communication convergence engineering
• /
• 제13권3호
• /
• pp.152-158
• /
• 2015

We consider the problem of designing independently operating local quantizers at nodes in distributed estimation systems, where many spatially distributed sensor nodes measure a parameter of interest, quantize these measurements, and send the quantized data to a fusion node, which conducts the parameter estimation. Motivated by the discussion that the estimation accuracy can be improved by using the quantized data with a high probability of occurrence, we propose an iterative algorithm with a simple design rule that produces quantizers by searching boundary values with an increased likelihood. We prove that this design rule generates a considerably reduced interval for finding the next boundary values, yielding a low design complexity. We demonstrate through extensive simulations that the proposed algorithm achieves a significant performance gain with respect to traditional quantizer designs. A comparison with the recently published novel algorithms further illustrates the benefit of the proposed technique in terms of performance and design complexity.

• 한국지능시스템학회논문지
• /
• 제11권8호
• /
• pp.689-694
• /
• 2001

Recently, several promising multiobjective evolutionary algorithm such as SPEA. NSGA-II, PESA, and SPEA2 have been developed. In this paper, we also propose a new multiobjective evolutionary algorithm that compares to them. In the algorithm proposed in this paper, we introduce a novel concept, “inherited age” and total algorithm is executed based on the inherited age concept. Also, we propose a new sharing algorithm, called objective classication sharing algorithm(OCSA) that can preserve the diversity of the population. We will show the superior performance of the proposed algorithm by comparing the proposed algorithm with other promising algorithms for the test functions.

• International Journal of Contents
• /
• 제7권4호
• /
• pp.30-34
• /
• 2011

A Leader is a Coordinator that supports a set of processes to cooperate a given task. This concept is used in several domains such as distributed systems, parallelism and cooperative support for cooperative work. In completely asynchronous systems, there is no solution for the election problem satisfying both of safety and liveness properties in asynchronous distributed systems. Therefore, to solve the election problem in those systems, one property should be weaker than the other property. If an election algorithm strengthens the safety property in sacrifice of liveness property, it would not nearly progress. But on the contrary, an election algorithm strengthening the liveness property in sacrifice of the safety property would have the high probability of violating the safety property. In this paper, we presents a safety strengthened Leader Election protocol with an unreliable failure detector and analyses it in terms of safety and liveness properties in asynchronous distributed systems.

• International Journal of Control, Automation, and Systems
• /
• 제6권2호
• /
• pp.282-287
• /
• 2008

In distributed mobile robot systems, autonomous robots accomplish complicated tasks through intelligent cooperation with each other. This paper presents behavior learning and online distributed evolution for cooperative behavior of a group of autonomous robots. Learning and evolution capabilities are essential for a group of autonomous robots to adapt to unstructured environments. Behavior learning finds an optimal state-action mapping of a robot for a given operating condition. In behavior learning, a Q-learning algorithm is modified to handle delayed rewards in the distributed robot systems. A group of robots implements cooperative behaviors through communication with other robots. Individual robots improve the state-action mapping through online evolution with the crossover operator based on the Q-values and their update frequencies. A cooperative material search problem demonstrated the effectiveness of the proposed behavior learning and online distributed evolution method for implementing cooperative behavior of a group of autonomous mobile robots.

• 산업경영시스템학회지
• /
• 제18권36호
• /
• pp.143-152
• /
• 1995

This paper determines the allocation of computers and data files to minimize the sum of processing and communication costs which occur in processing jobs at each node. The problem of optimally configuring a distributed computer system belongs to the class of NP-Complete problems and the object function of this paper is nonlinear function and is hard to solve. This paper seeks the solution of distributed processing system by Tabu Search. Firstly, it presents the method of generating the starting solution proper to the distributed processing system. Secondly, it develops the method of searching neighborhood solutions. Finally, it determines the Tabu restriction appropriate to the distributed processing system. According to the experimental results, this algorithm solves a sized problems in reasonable time and is effective in the convergence of the solution. The algorithm developed in this paper is also applicable to the general allocation problems of the distributed processing system.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제6권11호
• /
• pp.453-462
• /
• 2017

본 논문에서는 SDN (Software Defined Networking) 오픈소스 프레임워크인 오픈데이라이트(ODL, OpenDaylight) 컨트롤러 클러스터 환경에서 클러스터의 구조를 분석하며 고가용성(High availability)을 지원하는 컨트롤러 클러스터의 동작 방식을 다룬다. 또한 Raft 알고리즘의 리더 선정(Leader Election) 과정을 분석하고 효율적인 시스템 운용을 위한 Leader Election 과정의 개선 방안을 제안한다. 이와 함께 샤드(Shard) 리더와 샤드 팔로어의 성능차이를 제시하고, 기존과 제안 방식의 컨트롤러 클러스터의 성능을 비교 분석한다. 실험의 결과에 따르면 리더의 성능은 팔로어의 성능보다 좋으며 하나의 컨트롤러로 요청이 집중되어 전달될 때보다 분산된 컨트롤러로 요청이 전달될 때의 성능이 더 좋다. 따라서 제안 기법을 통하여 컨트롤러로의 요청을 분산함으로써 성능을 높일 수 있다.

• 한국통신학회논문지
• /
• 제35권4A호
• /
• pp.315-321
• /
• 2010

Ad-hoc 네트워크에서는 전력을 제어할 기지국의 부재로 시스템의 성능을 최적화 시키는 중앙 전력 제어가 불가능하여, 각 노드들은 독립적이고 자동적인 방식으로 분산형 전력 제어 알고리즘을 수행해야 한다. 기존의 분산형 전력 제어 알고리즘은 수신 신호의 SINR (signal to interference and noise ratio)에 따라 MCS (modulation and coding scheme) 레벨을 변화 시키는 adaptive bit loading operation을 고려하지 않기 때문에, 전체 throughput 향상에 제한이 있다. 본 논문에서는 Ad-hoc 환경에서 전체 throughput 을 향상시키고, outage 확률을 낮추기 위해 MCS 레벨에 따라 adaptive bit loading operation을 고려한 분산형 전력 제어 알고리즘을 제안한다. 컴퓨터 모의실험 결과 매우 큰 throughput 향상과 outage 확률 감소의 성능 향상을 확인할 수 있다.

• 한국콘텐츠학회논문지
• /
• 제16권7호
• /
• pp.667-680
• /
• 2016

하둡 맵리듀스와 같은 분산 컴퓨팅 플랫폼이 개발됨에 따라, 기존 단일 컴퓨터 상에서 수행되는 질의 처리 기법을 분산 컴퓨팅 환경에서 효율적으로 수행하는 것이 필요하다. 특히, 주어진 두 데이터 집합에서 유사도가 높은 모든 데이터 쌍을 탐색하는 유사 조인 질의를 분산 컴퓨팅 환경에서 수행하려는 연구가 있어 왔다. 그러나 분산 병렬 환경에서의 기존 유사 조인 질의처리 기법은 데이터 전송 비용만을 고려하기 때문에 클러스터 간에 비균등 연산 부하 분산의 문제점이 존재한다. 본 논문에서는 분산 컴퓨팅 환경에서 효율적인 유사 조인 처리를 위한 행렬 기반 부하 분산 알고리즘을 제안한다. 제안하는 알고리즘은 클러스터의 균등 부하 분산을 위해 행렬을 이용하여 예상되는 연산 부하를 측정하고 이에 따라 파티션을 생성한다. 아울러, 클러스터에서 질의 처리에 사용되지 않는 데이터를 필터링함으로서 연산 부하를 감소시킨다. 마지막으로 성능 평가를 통해 제안하는 알고리즘이 기존 기법에 비해 질의 처리 성능 측면에서 우수함을 보인다.

• 산업경영시스템학회지
• /
• 제41권4호
• /
• pp.179-188
• /
• 2018

Recently, blockchain technology has been recognized as one of the most important issues for the 4th Industrial Revolution which can be represented by Artificial Intelligence and Internet of Things. Cryptocurrency, named Bitcoin, was the first successful implementation of blockchain, and it triggered the emergence of various cryptocurrencies. In addition, blockchain technology has been applied to various applications such as finance, healthcare, manufacturing, logistics as well as public services. Distributed consensus algorithm is an essential component in blockchain, and it enables all nodes belonging to blockchain network to make an agreement, which means all nodes have the same information. For example, Bitcoin uses a consensus algorithm called Proof-of-Work (PoW) that gives possession of block generation based on the computational volume committed by nodes. However, energy consumption for block generation in PoW has drastically increased due to the growth of computational performance to prove the possession of block. Although many other distributed consensus algorithms including Proof-of-Stake are suggested, they have their own advantages and limitations, and new research works should be proposed to overcome these limitations. For doing this, above all things, we need to establish an evaluation method existing distributed consensus algorithms. Based on this motivation, in this work, we suggest and analyze assessment items by classifying them as efficiency and safety perspectives for investigating existing distributed consensus algorithms. Furthermore, we suggest new assessment criteria and their implementation methods, which can be used for a baseline for improving performance of existing distributed consensus algorithms and designing new consensus algorithm in future.