Earticle

Debugging multithreaded or parallel programs is a hard problem due to unintended non-deterministic results of the program, called data races. Previous race detection techniques have limitations of coverage for synchronization constructs in detecting first races for parallel programs with random synchronizations. We present, in this paper, an efficient two-pass algorithm to detect the first races in a particular execution of parallel program, by eliminating independent accesses to make a performance gain and candidate accesses, and then analyzing the traced information with candidate accesses by replaying the program. Therefore, our technique makes the race detection more practical in debugging parallel programs which have random synchronizations.

Hadoop–a popular open-source implementation of MapReduce is widely used for the analysis of large datasets. The current Hadoop implementation assumes that computing nodes in a cluster are homogeneous in nature. In this paper we evaluate performance of Hadoop Platform and Oracle for Distributed Parallel Processing in large datasets. For evaluation, we implement a prototype of a virtual datacenter using distributed and parallel computing technology. The purpose of this paper is to reduce datacenter implementation cost using commodity hardware and provide high performance. Hadoop is installed on a commodity Linux cluster the distributed processing of large data sets across clusters of computers using distributed and parallel computing architecture. This paper also helps to explain about some new technology and framework which are open source; that can easily utilize those technologies for our complex data analysis which resembling structured, semi structured and non-structured data. Here we tried to demonstrate a performance comparison by executing some queries between distributed parallel computing system and traditional single computing system. For the simulation of the infrastructure Hadoop cluster has been used for distributed parallel processing and Oracle 11g is used for traditional single processing system. We prepare three virtual host for Hadoop cluster and a high-end hardware for Oracle 11g.

Generally, multidimensional data such as image and spatial data require large amount of storage space. There is a limit to store and manage those large amounts of data in single workstation. If we manage the data on parallel computing environment which is being actively researched these days, we can get highly improved performance. In this paper, we propose a parallel multidimensional index structure that exploits the parallelism of the parallel computing environment. The proposed index structure is nP(processor)-nxmD(disk) architecture which is the hybrid type of nP-nD and 1P-nD. Its node structure in-creases fan-out and reduces the height of an index. Also, a range search algorithm that maximizes I/O parallelism is devised, and it is applied to k-nearest neighbor queries. Through various experiments, it is shown that the proposed method outperforms other parallel index structures.

Real-time multimedia streaming over the Internet is rapidly increasing with the popularity of user-created contents, Web 2.0 trends, and P2P (peer-to-peer) delivery support. While many homes today are broadband-enabled, the quality of experience (QoE) of a user is still limited due to frequent interruption of media playout. The vulnerability of TCP (transmission control protocol), the popular transport-layer protocol for streaming in practice, to the packet losses, retransmissions, and timeouts makes it hard to deliver a timely and persistent flow of packets for online multimedia contents. This paper presents TCP-real-time online multimedia environment (ROME), a novel transport-layer framework that allows the establishment and coordination of multiple many-to-one TCP connections. Between one client with multiple home addresses and multiple co-located or distributed servers, TCP-ROME increases the total throughput by aggregating the resources of multiple TCP connections. It also overcomes the bandwidth fluctuations of network bottlenecks by dynamically coordinating the streams of contents from multiple servers and by adapting the streaming rate of all connections to match the bandwidth requirement of the target video.

본 연구는 금속제품 생산 프로세스의 처리능력을 결정하는 주요 공정인 열처리 공정의 생산계획을 수립하는 것이다. 열처리로 전체의 작업완료시간(makespan)을 최소화하기 위한 유전알고리즘을 제시하는 것을 본 연구의 목표로 한다. 일반적인 유전알고리즘의 각 단계에서 해의 가능성을 만족하기 위해 ‘재할당 절차’를 통해 알고리즘의 적용을 유효하게 하였고, SPT 규칙을 기초로 한 FFD 방법론을 활용하여 각 열처리로 별로 세부 일정계획을 수립하였다. 이러한 절차를 통해 효과적인 해를 찾을 수 있도록 하였으며, 유전알고리즘 해와 비교를 위해 하한값을 제시하였다. 제시된 유전알고리즘의 성능을 분석하기 위해 성능평가를 실시하였다.

This paper considers a batch scheduling problem for heat-treatment process with parallel machine environments. To derive efficient and effective solutions for the targeted problem, this paper considers the genetic algorithm to minimize the makespan objective measure. To satisfy the feasibility of the solutions provided by the GA algorithm, this paper derives some solution properties and procedures, such as 'relocation procedure' etc. In addition, using SPT rule based FFD methodology, detailed schedules are established at each heat treatment. To show the effectiveness of the solutions by GA algorithm, some lower bounds are also derived. The computational experiments are conducted to test the performance of the derived GA algorithm.

네트워크를 통하여 서로 연결된 컴퓨팅 자원들의 집합을 분산 시스템이라고 정의할 수 있다. 최적화 문제 영역에서 가장 중요한 해결 기법 중에 하나인 병렬 유전자 알고리즘은 분산 시스템을 기반으로 하고 있다. 인터넷과 이동 컴퓨팅과 같은 동적 네트워크 환경 하에서 네트워크의 상태는 가변적으로 변할 수 있어 기존의 병렬 유전자 알고리즘을 분산 시스템에서 최적화 문제를 해결하기 위하여 그대로 사용하기에는 비효율적이다. 본 논문에서는 동적 네트워크 환경 하에서 분산 에이전트를 사용하여 병렬 유전자 알고리즘을 효율적으로 사용할 수 있는 기법을 제시한다.

Distributed Systems can be defined as set of computing resources connected by computer network. One of the most significant techniques in optimization problem domains is parallel genetic algorithms, which are based on distributed systems. Since the status of dynamic network environments such as Internet and mobile computing. can be changed continually, it must not be efficient on the dynamic environments to solve an optimization problem using previous parallel genetic algorithms themselves. In this paper, we propose the effective technique, in which the parallel genetic algorithm can be used efficiently on the dynamic network environments.

병렬 스케줄링의 목적은 다중프로세서 시스템 환경에서 병렬성을 가진 응용프로그램에 대해 최소의 동기화 오버헤드와 부하균등(load balance)을 달성하도록 스케줄링을 수행하는데 있다. 프로세서들이 병렬 반복(iteration)을 실행하기 위해서는 메모리로부터 반복들에 대한 chunk를 계산하고 할당받게 된다. 이때, 전역 메모리의 상호 배타적인 빈번한 접근으로 많은 스케쥴링 오버헤드 및 병목현상이 발생된다. 또한, 프로세서에게 할당된 chunk내 병렬 반복들의 분포가 서로 상이한 경우,각 chunk의 실행시간이 서로 달라 부하불균등의 원인이 되어 결과적으로 전체 스케쥴링 성능에 나쁜 영향을 준다. 따라서, 최소의 스케줄링 오버 헤드와 부하균등을 달성하기 위해 본 논문에 서는 기존의 방법들에서 문제점들을 도출하고, 자료의 국부성과 프로세서 동족성(affinity)을 고려한 병렬 루프 스케줄링 알고리즘을 제안한다.

The purpose of a parallel scheduling under a multiprocessor environment is to carry out the scheduling with the minimum synchronization overhead, and to perform load balance for a parallel application program. The processors calculate the chunk of iteration and are allocated to carry out the parallel iteration. At this time, it frequently accesses mutually exclusive global memory so that there are a lot of scheduling overhead and bottleneck imposed. And also, when the distribution of the parallel iteration in the allocated chunk to the processor is different, the different execution time of each chunk causes the load imbalance and badly affects the capability of the all scheduling. In the paper. we investigate the problems on the conventional algorithms in order to achieve the minimum scheduling overhead and load balance. we then present a new parallel loop scheduling algorithm, considering the locality of the data and processor affinity.

강화 학습을 통한 실내 자율주행을 위해 가상 환경과 실제 환경에서 학습을 병행하는 방법을 제안한다. 실제 환경에서만 학습을 진행했을 경우 80시간 정도의 소요 시간이 필요하지만, 실제 환경과 가상 환경을 병행하며 학습을 진행했을 경우 50시간의 소요 시간이 필요하다. 가상 환경과 실제 환경에서 학습을 병행하면서 빠른 학습으로 다양한 실험을 거쳐 최적화된 파라미터를 얻을 수 있는 이점이 있다. 실내복도 이미지를 이용하여 가상 환경을 구성한 후 데스크톱으로 선행학습을 진행하였고 실제 환경에서의 학습은 Jetson Xavier를 기반으로 다양한 센서와 연결하여 학습을 진행하였다. 또한, 실내복도 환경의 반복되는 텍스처에 따른 정확도 문제를 해결하기 위해 복도 벽의 아랫선을 강조하는 특징점 검출을 학습하여 복도 벽 객체를 판단하고 정확도를 높일 수 있었다. 학습을 진행할수록 실험 차량은 실내복도 환경에서 복도 중앙을 기준으로 주행하며 평균 70회의 조향명령을 통해 움직인다.

We propose a method that combines learning in a virtual environment and a real environment for indoor autonomous driving through reinforcement learning. In case of learning only in the real environment, it takes about 80 hours, but in case of learning in both the real and virtual environments, it takes 40 hours. There is an advantage in that it is possible to obtain optimized parameters through various experiments through fast learning while learning in a virtual environment and a real environment in parallel. After configuring a virtual environment using indoor hallway images, prior learning was carried out on the desktop, and learning in the real environment was conducted by connecting various sensors based on Jetson Xavier. In addition, in order to solve the accuracy problem according to the repeated texture of the indoor corridor environment, it was possible to determine the corridor wall object and increase the accuracy by learning the feature point detection that emphasizes the lower line of the corridor wall. As the learning progresses, the experimental vehicle drives based on the center of the corridor in an indoor corridor environment and moves through an average of 70 steering commands.

노후·위험 주거환경 문제는 개인의 생활 불편을 넘어 건강과 삶의 질 격차를 심화시키는 구조적 사회문 제로 부각되고 있다. 이에 따라 주거환경 개선사업을 통한 정책적 개입이 공공정책과 사회혁신 영역에서 중요한 대응 전략으로 대두되고 있으나, 이러한 개입이 거주자의 삶의 질 개선으로 이어지는 메커니즘에 대한 구조적 설명은 여전히 제한적이다. 본 연구는 주거 스트레스 이론과 집단효능감 이론을 바탕으로 주 거환경 개선이 주거 스트레스 감소와 집단효능감 증진을 통해 거주자의 삶의 질 개선으로 이어지는 구조적 경로를 통계적으로 검증했다. 또한 수혜자가 인식하는 개선사업의 가치가 이러한 경로를 조절하는 조절된 매개효과를 탐색했다. 이를 위해 2025년에 서울 소재의 모(某) 기초자치단체가 수행한 위험거처 주거환경 개선사업의 수혜자 104명을 대상으로 설문조사를 실시하고, PROCESS macro를 활용하여 병렬 및 조절된 매개모형을 분석했 다. 분석 결과, 주거환경 개선은 삶의 질에 유의한 정(+)의 영향을 미쳤으며, 주거 스트레스 감소와 집단효 능감 증진을 통한 부분 매개효과가 확인됐다. 특히 집단효능감을 경유한 간접효과가 상대적으로 크게 나타났다. 또한 지각된 가치의 조절 분석에서는 주거 스트레스 완화 경로에서 지각된 가치 수준에 따라 효과가 조건부로 달라지며, 특히 높은 수준에서 효과가 약화되는 경향이 관찰되었다. 집단효능감 경로에서는 완전 한 조절된 매개효과가 통계적으로 지지되지 않았다. 본 연구는 주거환경 개선 효과가 개인의 심리적 스트레스 감소와 함께, 개인이 자각하는 공동체 역량 수준 의 증진을 통해 다층적으로 작동함을 실증적으로 확인했다. 이는 주거복지 정책에 있어서 주거환경의 물리 적 개선과 함께 수혜자의 가치 인식과 사회적 관계를 통합적으로 고려할 필요성을 시사한다.

The problem of deteriorated and hazardous housing environments has emerged as a structural social issue that extends beyond individual inconvenience, exacerbating disparities in health and quality of life. Consequently, policy interventions through housing environment improvement programs have gained prominence as a key response strategy in public policy and social innovation. However, structural explanations of the mechanisms through which such interventions translate into improvements in residents’ quality of life remain limited. This study draws on Dwelling Stress Theory and Collective Efficacy Theory to statistically verify the structural pathways through which housing environment improvement leads to enhanced quality of life via reductions in dwelling stress and increases in collective efficacy. It further explores whether beneficiaries’ perceived value of the intervention moderates these pathways through a moderated mediation framework. Survey data were collected in 2025 from 104 participants (79 beneficiaries and 25 comparison group) of a housing environment improvement project implemented by a district-level local government in Seoul, and parallel and moderated mediation models were estimated using the PROCESS macro. The results indicate that housing environment improvement had a significant positive effect on quality of life, with both dwelling stress reduction and collective efficacy enhancement confirmed as partial mediators. Notably, the indirect pathway through collective efficacy was relatively larger than that through dwelling stress. In the moderated mediation analysis, it was observed that the effect of housing stress reduction on quality of life was conditional upon the level of perceived value; specifically, this effect tended to weaken at higher levels of perceived value. In contrast, a fully supported moderated mediation effect through the collective efficacy pathway was not statistically confirmed. This study empirically demonstrates that the effects of housing environment improvement operate through a multi-layered mechanism encompassing both individual-level psychological stress reduction and community-level enhancement of perceived collective capacity. The findings suggest that housing welfare policies should move beyond a sole focus on physical improvements to integrate beneficiaries’ value perceptions and social relational dynamics into more comprehensive and context-sensitive intervention designs.