• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.441-452
• /
• 2021

Federated learning provides an efficient integrated model for distributed data, allowing the local training of different data. Meanwhile, the goal of multi-task learning is to simultaneously establish models for multiple related tasks, and to obtain the underlying main structure. However, traditional federated multi-task learning models not only have strict requirements for the data distribution, but also demand large amounts of calculation and have slow convergence, which hindered their promotion in many fields. In our work, we apply the rank constraint on weight vectors of the multi-task learning model to adaptively adjust the task's similarity learning, according to the distribution of federal node data. The proposed model has a general framework for solving optimal solutions, which can be used to deal with various data types. Experiments show that our model has achieved the best results in different dataset. Notably, our model can still obtain stable results in datasets with large distribution differences. In addition, compared with traditional federated multi-task learning models, our algorithm is able to converge on a local optimal solution within limited training iterations.

• Journal of Korean Society for Geospatial Information Science
• /
• v.19 no.3
• /
• pp.23-32
• /
• 2011

This research uses a distributed model, Vflo which can devide subwater shed into square grids and interpret diverse topographic elements which are obtained through GIS processing. To use the distributed model, soil wetness information was extracted through Tasseled Cap transformation from LANDSAT 7 $ETM^+$ satellite data and then they were applied to each cell of the test area, unlike previous studies in which have applied average soil condition of river basin uniformly regardless of space-difference in subwater shed. As a resut of the research, it was ascertained the spatial change of soil wetness is suited to the distributed model in a subwater shed. In addition, we derived out a relation between soil wetness of image collection time and 10 days-preceded rainfall and improved the feasibility of weights obtained by the relation equation.

• The KIPS Transactions:PartA
• /
• v.9A no.1
• /
• pp.45-52
• /
• 2002

This paper proposes the fuzzy closed RCMP queueing network model using fuzzy set theory for the performance evaluation of centralized distributed processing system with ambiguous system factors in the network environments. This model can derive the measures for system performances such as the job spending time, the system throughput, average job number and server utilizations using fuzzy mean value analysis which can process the fuzzy factors. Computer simulation has been performed centralized distributed system with fuzzy service requirement time for verifying the effectiveness of derived equations of performance evaluation according to the numbers of clients, and the results were analyzed. The proposed model provides more and flexible realistic than performance evaluation of conventional method when we evaluated system performance with ambiguous factors.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.5
• /
• pp.85-91
• /
• 2009

The centralized conference architecture has a restriction in scalability due to the performance reduction as the number of conference participants increases. To solve this problem several distributed conference architectures have been studied recently. In these architectures new conference servers are added dynamically to the conference environment. In this paper, We have proposed a new conference information data model which can be used in these distributed conference architectures. In our newly proposed conference information data model. several components has been added for exchanging conference information between primary conference server and multiple secondary conference servers. We also proposed a procedure of conference information exchange between these conference servers. And the management of conference informations and SIP(Session Initiation Protocol) notifications to the total conference participants can be processed distributedly with these conference servers, therefore the load to the primary conference server can be decreased by using this method. The performance of our proposed model has been evaluated by experiments.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.2
• /
• pp.659-679
• /
• 2015

Distributed applications are composed of multiple nodes, which exchange information with individual nodes through message passing. Compared with traditional applications, distributed applications have more complex behavior patterns because a large number of interactions and concurrent behaviors exist among their distributed nodes. Thus, it is difficult to detect anomalous behaviors and determine the location and scope of abnormal nodes, and some attacks and misuse cannot be detected. To address this problem, we introduce a method for detecting anomalous behaviors based on process algebra. We specify the architecture of the behavior detection model and the detection algorithm. The anomalous behavior detection and analysis demonstrate that our method is a good discriminator between normal and anomalous behavior characteristics of distributed applications. Performance evaluation shows that the proposed method enhances efficiency without security degradation.

• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 2007.02a
• /
• pp.131-138
• /
• 2007

Distributed Group Decision Support System (GDSS) is in the stage between exploration and implementation, there is not unified constructing model. As computer software and hardware, network technique develop, especially the development of object-oriented programming, distributed process, and artificial intelligence, this makes it possible the practical and valid implementation of distributed GDSS. With a view of emphasizing and solving process-supporting, this article discusses how to use the key technologies of network, distributed process, artificial intelligence and man-machine mutual interface, to implement more adaptable, more flexible, and more valid GDSS than before.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.4
• /
• pp.633-641
• /
• 2018

The risk of disasters, such as floods and drought, has increased. Reliable hydrological data is important for analyzing the water resource and designing hydraulic structure to manage these risks. The Yongdam Guryang river catchment located in the central of Korea is the research catchment of K-water and UNESCO IHP, and the hydrological data, such as rainfall, runoff, evapotranspiration, etc. has been observed at the catchment. The aim of this study was to assess the runoff characteristics of the small mountainous catchment of Korea based on the observed hydrological data, and the Probability Distributed Model was applied as the Rainfall-Runoff Model at the Yongdam Guryang river catchment. The hydrological data was divided into the wet period from June to September and dry period from October to May according to data analysis. The runoff ratio was 0.27~0.41 in the wet period and 0.30~0.45 in the dry period. The calibration result by the Probability Distributed Model showed a difference in the calibrated model parameters according to the periods. In addition, the model simulated the runoff accurately except for the dry period of 2015, and the result revealed the applicability of the PDM. This study showed the runoff characteristics of the small mountainous catchment by dividing the hydrological data into dry and wet periods.

• Earthquakes and Structures
• /
• v.5 no.2
• /
• pp.239-259
• /
• 2013

Typically, beams that form part of structural systems are subjected to vertical distributed loading along their length. Distributed loading affects moment and shear distribution, and consequently spread of inelasticity, along the beam length. However, the finite element models developed so far for seismic analysis of frame structures either ignore the effect of vertical distributed loading on spread of inelasticity or consider it in an approximate manner. In this paper, a beam-type finite element is developed, which is capable of considering accurately the effect of uniform distributed loading on spreading of inelastic deformations along the beam length. The proposed model consists of two gradual spread inelasticity sub-elements accounting explicitly for inelastic flexural and shear response. Following this approach, the effect of distributed loading on spreading of inelastic flexural and shear deformations is properly taken into account. The finite element is implemented in the seismic analysis of plane frame structures with beam members controlled either by flexure or shear. It is shown that to obtain accurate results the influence of distributed beam loading on spreading of inelastic deformations should be taken into account in the inelastic seismic analysis of frame structures.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2005.05a
• /
• pp.1033-1039
• /
• 2005

Parallel and distributed simulation is concerned with the efficient execution of large-scale discrete event simulation models on multiprocessors and distributed platforms. After the development of WWW, many efforts in the parallel and distributed simulation have been made for modeling, particularly building simulation languages and creating model libraries that can be assembled and executed over WWW. However, web-based parallel and distributed simulation is restricted by heterogeneous computing environments. Recently, the advent of XML and web services technology has made these efforts enter upon a new phase. Especially, the web services as a distributed information technology have demonstrated powerful capabilities for scalable interoperation of heterogeneous systems. This paper aims to develop and evaluate the parallel and distributed simulation using the web services technology. In particular, a prototype multi-pass simulation framework is implemented using Java-based web services technology. It focuses on the efficiency of multi-pass simulation used for optimization through the distribution of simulation replication to several simulation service providers. The development of parallel and distributed simulation using web services will help solve efficiently large-scale problems and also guarantee interoperability among heterogeneous networked systems.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.9 no.3
• /
• pp.25-32
• /
• 2006

Weapon systems composed of several subsystems execute various engagement missions in distributed combat environments in cooperation with a large number of subordinate/adjacent weapon systems as well as higher echelons through tactical data links. Such distributed weapon systems require distributed real-time simulation test beds to integrate and test their operational software, analyze their performance and effects of cooperated engagement, and validate their requirement specifications. These demands present significant challenges in terms of real-time constraints, time synchronization, complexity and development cost of an engagement simulation test bed, thus necessitate the use of high-performance distributed real-time simulation architectures, and modeling and simulation techniques. In this paper, in order to meet these demands, we presented a distributed real-time simulation system based on High Level Architecture(HLA) and Discrete Event System Specification(DEVS). We validated its performance by using it as a test bed for developing the Engagement Control System(ECS) of a surface-to-air missile system. The proposed technique can be employed to design a prototype or model of engagement-level distributed real-time simulation systems.