-
公开(公告)号：KR102165862B1
公开(公告)日：2020-10-14
申请号：KR1020190089187
申请日：2019-07-23
Applicant: 성균관대학교산학협력단
IPC: H04L12/715 , H04W84/18 , H04L12/801 , H04L12/751
-
公开(公告)号：KR1020160114241A
公开(公告)日：2016-10-05
申请号：KR1020150040268
申请日：2015-03-23
Applicant: 성균관대학교산학협력단
IPC: G06F17/30
Abstract: 본발명의실시예에따른컴퓨터를이용한데이터마이닝연관규칙생성방법은, 컴퓨터가, 주어진입력데이터셋으로부터후보연관규칙들을추출하는단계, 후보연관규칙들에기초한후보연관규칙행렬을잠재의미론적분석기법을통해분석하여각 아이템과이벤트플래그사이의유사도값들을산출하는단계및 유사도값에기초하여감축된연관규칙을생성하는단계를포함할수 있다.
-
公开(公告)号：KR101644750B1
公开(公告)日：2016-08-16
申请号：KR1020150079932
申请日：2015-06-05
Applicant: 성균관대학교산학협력단
IPC: H04L12/24
Abstract: 본발명은센서노드의잔류에너지데이터와전체센서노드의평균잔류에너지데이터를비교하여센서노드의활성모드를제어하고, 센서노드자체에서중복된센싱데이터를제어하여센서노드의에너지효율을향상시킨센서노드와서버, 이를포함하는무선센서네트워크시스템및 그운영방법에관한것이다. 본발명의일 실시예에따른무선센서네트워크시스템은, 센싱데이터및 잔류에너지데이터를생성하고, 싱크노드로센싱데이터와잔류에너지데이터를전송하는센서노드; 센서노드로부터전송받은센싱데이터및 잔류에너지데이터를서버로전송하는싱크노드; 및, 싱크노드로부터센서노드의센싱데이터와잔류에너지데이터를수신하고, 수신된센서노드의잔류에너지데이터를이용하여센서노드의동작모드를결정하고, 결정된동작모드에대한제어메시지를생성하는서버를포함한다.
Abstract translation: 本发明涉及传感器节点，服务器，包括传感器节点和服务器的无线传感器网络系统，以及用于管理无线传感器网络系统的方法，其中传感器节点的激活模式通过比较 具有所有传感器节点的平均剩余能量数据的传感器节点，并且通过在传感器节点的级别控制重复的感测数据来提高传感器节点的能量效率。 根据本发明的实施例，无线传感器网络系统包括：生成感测数据和剩余能量数据的传感器节点，并将感测数据和剩余能量数据传送到汇聚节点; 将从传感器节点接收的感测数据和剩余能量数据发送到服务器的汇聚节点; 并且从汇聚节点接收传感器节点的感测数据和剩余能量数据的服务器，通过使用接收到的传感器节点的剩余能量数据来确定传感器节点的操作模式，并且生成关于确定的操作模式的控制消息。
-
34.发明授权
公开(公告)号：KR101639947B1
公开(公告)日：2016-07-15
申请号：KR1020150052368
申请日：2015-04-14
Applicant: 성균관대학교산학협력단
Abstract: 본발명은하둡선점데드라인제약스케줄링방법및 그방법을수행하는컴퓨터프로그램과, 그프로그램이기록된매체에관한것으로, 하둡시스템에적용되는스케줄링방법에있어서, (a) 스케줄러가큐(Queue)를스캐닝하여다수의잡(Job)에대한정보를획득하고초기화하는단계; (b) 스케줄러가큐(Queue)를스캔하여사용가능한슬롯을찾아스케줄링하는단계; (c) 스케줄러가실행중인잡의실행완료시간까지의잔여시간을산출하고, 우선권이부여되고상기우선권의순서에따라새로운잡의상기잔여시간과데드라인의기간과비교하여상기현재실행중인잡의슬롯에상기새로운잡의선점여부를결정하는단계; 및 (d) 스케줄러가상기선점여부의결정에따라잡에게슬롯을부여하여할당하는단계를포한한다.이와같은본 발명은, 하둡환경하에서슬롯들의효과적인사용에의해데드라인아래수행되는잡들의수를최대화하고, 선점을지원하고, 선점의오버헤드를줄이기위한선점결정기준을개선하여성능을향상시킬수 있는하둡선점데드라인제약스케줄링방법을제공한다.
Abstract translation: 本发明涉及一种Hadoop抢占期限约束调度方法，执行该方法的计算机程序以及记录该程序的记录介质。 应用于Hadoop系统的调度方法包括以下步骤：（a）通过调度器获取关于多个作业的信息并通过扫描队列来执行初始化; （b）通过调度器扫描队列来查找和调度可用时隙; （c）通过计算当前执行的作业的执行完成时间的剩余时间来确定是否通过新作业抢占当前执行的作业的时隙，并将新作业的剩余时间和 按照优先顺序排列到最后期限; 和（d）通过调度器根据抢占状态的确定给该时隙给该作业。 如上所述的本发明提供了能够通过有效利用Hadoop环境中的时隙来最大化在截止日期之前执行的作业的数量的Hadoop抢先期限约束调度方法，并支持抢占，并且提高了用于减少开销的抢占确定标准 的抢占，从而提高业绩。
-
35.发明公开
公开(公告)号：KR1020160011776A
公开(公告)日：2016-02-02
申请号：KR1020140092752
申请日：2014-07-22
Applicant: 성균관대학교산학협력단
Abstract: 본발명의실시예에따른다중의사결정문제해법도출방법은, 의사결정에참여하는 K 명의의사결정자들이 n 개의대안들을 m 개의기준들에따라평가하는언어적변수들로구성된의사결정행렬을의사결정자마다생성하는단계, 정규화된의사결정행렬들을구하는단계, 각기준에서의엔트로피에기초하여각 기준에대한가중치를연산하는단계, 어느한 기준의다른기준에대한상대적중요도를산출하여생성되는판단행렬의각 행값들에기초하여연산되는상대적가중치들을생성하는단계, 각기준의가중치와각 기준의상대적가중치를곱한값들을정규화한포괄적가중치들을연산하고, 연산된포괄적가중치들과정규화회색의사결정행렬을이용하여가중정규화회색의사결정행렬을생성하는단계, 긍정적이상해법과부정적이상해법및 상기가중정규화의사결정행렬을이용하여, 회색상관계수및 회색상관지수를생성하고, 그룹분리척도를연산하는단계, 그룹분리척도에기초하여결정된상대적근접도들의크기순서에따라대안들의순위를결정하는단계를포함할수 있다.
Abstract translation: 根据本发明的一个实施例，一种用于导出多决策问题的解决方案的方法包括以下步骤：通过参与决策的K个决策者生成由语言变量组成的决策矩阵，该决策矩阵评估 根据每个决策者的参考数量，n个替代方案; 获得归一化决策矩阵; 基于每个参考中的熵的每个参考的操作权重值; 生成基于通过计算对于任何一个参考的其他参考的相对重要性而产生的确定矩阵的每个列值来操作的相对权重值; 操作归一化综合权重值，其乘以每个参考的权重值和每个参考的相对权重值，并且通过使用所操作的综合权重值和归一化灰色决策矩阵来生成加权归一化灰色决策矩阵; 生成灰色相关系数和灰色相关指标，并通过使用正的理想解，负理想解和加权归一化决策矩阵来操作组分离度量; 以及根据基于组分离测量确定的相对接近度的大小顺序来确定替代方案的排名。
-
Patent Agency Ranking
公开(公告)号：KR1020140080773A
公开(公告)日：2014-07-01
申请号：KR1020120148207
申请日：2012-12-18
Applicant: 성균관대학교산학협력단
CPC classification number: G06F17/16
Abstract: A method for obtaining solutions to make a decision multiple-opinions comprises the steps of (a) creating an n x m opinion decision matrix including language parameters used to evaluate n alternatives made by K decision makers participating in making a decision of the opinion according to m criteria; (b) normalizing an interval gray number corresponding to the language parameters for each opinion decision matrix of each decision maker to find the normalized opinion decision matrix; (c) calculating a weighted value for each criterion based on the variance at each criterion with respect to the normalized opinion decision matrix of each decision maker, and creating the weighted and normalized decision matrix by using the calculated weighted value; (d) deciding a positive ideal solution and a negative ideal solution from the weighted and normalized opinion decision matrix of each decision maker; (e) calculating a gray correlation coefficient with each decision maker by using the positive and negative ideal solutions and the weighted and normalized opinion decision matrix; (f) calculating each gray correlation index from the gray correlation coefficient for each alterative; (g) calculating a group division criterion for each alterative with respect to the gray correlation index of the opinion decision maker; (h) determining a proximity degree based on the group separation criterion for each alternative; and determining a level of the alternatives according to sizes of proximity degrees.
Abstract translation: 一种用于获得决策多重意见的解决方案的方法包括以下步骤：（a）创建一个nxm意见决策矩阵，其包括用于评估参与根据m个标准来决定该意见的K个决策者所做出的n个备选项的语言参数 ; （b）归一化与每个决策者的每个意见决策矩阵相对应的语言参数的间隔灰度数，以找到归一化的意见决策矩阵; （c）基于每个标准的方差相对于每个决策者的归一化意见决策矩阵来计算每个标准的加权值，并通过使用计算的加权值创建加权和归一化的判定矩阵; （d）从每个决策者的加权和归一化意见决策矩阵中确定积极的理想解和负理想解; （e）通过使用正和负理想解和加权和归一化意见决策矩阵来计算每个决策者的灰色相关系数; （f）从每个替代的灰色相关系数计算每个灰色相关指数; （g）根据意见决策者的灰色相关指数计算每个替代品的组划分标准; （h）基于每个替代物的组分离标准确定接近程度; 以及根据接近程度的大小来确定替代的水平。
-
公开(公告)号：KR1020130070758A
公开(公告)日：2013-06-28
申请号：KR1020110137940
申请日：2011-12-20
Applicant: 성균관대학교산학협력단
IPC: H04N21/63 , H04N21/6405
CPC classification number: H04N21/6405 , H04N21/2385
Abstract: PURPOSE: A VOD(Video On Demand) batching method combining static multi-casting and dynamic multi-casting is provided to combine the static multi-casting and the dynamic multi-casting by using interval time at transmission time, thereby effectively reducing the amount of use of channels. CONSTITUTION: When a segment number, which is requested in a dynamic broadcast group, is lesser than a next segment number which is transmitted to a static broadcast group, a system confirms continuous reproduction. When the segment number is greater than or equal to the next segment number, the system transmits the segment number to the dynamic broadcast group. When the continuous reproduction is possible, the system transmits the same segment to the dynamic broadcast group and the static broadcast group. When the continuous reproduction is impossible, the system transmits the segment number to the dynamic broadcast group. [Reference numerals] (AA) Start; (BB) Detect a segment(Rd) which is the most requested in a dynamic broadcast group; (CC) Confirm a next segment(Ns) transmitted from a static broadcast group; (DD) Whether or not continuous reproduction; (EE,GG) Transmit the segment(Rd) requested in the dynamic broadcast group; (FF) Transmit the same segment to the static broadcast group and the dynamic broadcast group
Abstract translation: 目的：提供一种组合静态多层铸造和动态多铸造的VOD（视频点播）批量化方法，通过使用传输时间间隔时间来组合静态多播和动态多播，从而有效减少 使用渠道。 构成：当在动态广播组中请求的段号小于被发送到静态广播组的下一个段号时，系统确认连续再现。 当段号大于或等于下一段号时，系统将段号发送到动态广播组。 当连续再现可能时，系统将相同的片段发送到动态广播组和静态广播组。 当连续再现不可能时，系统将段号发送到动态广播组。 （附图标记）（AA）开始; （BB）检测动态广播组中最多请求的段（Rd）; （CC）确认从静态广播组发送的下一个段（Ns）; （DD）是否连续繁殖; （EE，GG）发送动态广播组中请求的段（Rd）; （FF）将相同的段发送到静态广播组和动态广播组
-
公开(公告)号：KR1020130066788A
公开(公告)日：2013-06-21
申请号：KR1020110133449
申请日：2011-12-13
Applicant: 성균관대학교산학협력단
CPC classification number: H04L67/327 , G06F17/15
Abstract: PURPOSE: A context awareness system based on hierarchical context modeling is provided to use a mahalanobis distance for dividing a level of the hierarchical context modeling, thereby increasing the accuracy of context inference. CONSTITUTION: A context collection agent unit(110) collects lower level context information from a sensor device. A context awareness agent unit includes a context modeling unit(121) and a context inference unit(122). The context modeling unit models the context information by receiving the context information and using multi-hierarchical context modeling. The context inference unit infers the context information in a context model. An agent platform unit(130) communicates with the context collection agent unit and the context awareness agent unit by using a message transmission protocol. [Reference numerals] (100) Hierarchical context modeling based context awareness system; (110) Context collection agent unit; (120) Context recognition agent unit; (121) Context modeling unit; (122) Context inference unit; (130) Agent platform unit
Abstract translation: 目的：提供基于层次上下文建模的上下文感知系统，以使用马哈拉诺比斯距离来分割层次上下文建模的级别，从而增加上下文推断的准确性。 构成：上下文收集代理单元（110）从传感器设备收集较低级别的上下文信息。 上下文感知单元包括上下文建模单元（121）和上下文推理单元（122）。 上下文建模单元通过接收上下文信息并使用多层次上下文建模来模拟上下文信息。 上下文推理单元推断上下文模型中的上下文信息。 代理平台单元（130）通过使用消息传输协议与上下文收集代理单元和上下文感知代理单元进行通信。 （附图标记）（100）基于上下文建模的层次上下文建模系统; （110）上下文收集代理单元; （120）上下文识别单元; （121）上下文建模单元; （122）上下文推理单元; （130）代理平台单元
-
公开(公告)号：KR1020130066787A
公开(公告)日：2013-06-21
申请号：KR1020110133448
申请日：2011-12-13
Applicant: 성균관대학교산학협력단
IPC: H04W84/18
CPC classification number: H04W84/18 , H04L41/0823
Abstract: PURPOSE: A balanced chain topology generation method in a wireless sensor network is provided to reduce the weighting of a network by constructing a chain based on a minimum spanning tree algorithm. CONSTITUTION: A wireless sensor network is set as a square based on a base station(S1). The square network is divided into the sub-areas of a square type including the same size according to the number of nodes. A bridge node is selected for the connection of the sub area in order to form a network chain in the divided sub-area(S2). A node connection chain is configured in each sub area to connect the bridge node(S3). [Reference numerals] (S1) Step of setting a wireless sensor network as a square based on a BS, and dividing the square network into n^2 sub-areas of a square type including the same size according to the number of nodes; (S2) Step of selecting a bridge node which connects the sub-areas in order to form the entire chain of a network in the sub-areas divided in the step (S1); (S3) (S3) step of configuring a chain which connects nodes in each sub-area and forming the entire chain by connecting the bridge node; (S4) Step of re-configuring the configuration of the chain in a sub-area to which a failure node is belonged when the failure node is generated
Abstract translation: 目的：提供无线传感器网络中的平衡链拓扑生成方法，通过构建基于最小生成树算法的链来减少网络的权重。 构成：将无线传感器网络设置为基于基站的平方（S1）。 方形网络根据节点数量划分为包括相同大小的正方形子区域。 为了在分割子区域（S2）中形成网络链，选择用于子区域的连接的网桥节点。 在每个子区域中配置节点连接链以连接网桥节点（S3）。 （S1）将无线传感器网络设为基于BS的平方的步骤，根据节点数将平方网划分为包括相同大小的正方形的n ^ 2个子区域; （S2）在步骤（S1）中划分的子区域中选择连接子区域以形成网络的整个链路的网桥节点的步骤; （S3）（S3）构成通过连接桥节点来连接各子区域中的节点并形成整个链的链的步骤; （S4）当生成故障节点时，在故障节点所属的子区域中重新配置链路的配置的步骤
-
公开(公告)号：KR101137570B1
公开(公告)日：2012-04-20
申请号：KR1020090120669
申请日：2009-12-07
Applicant: 성균관대학교산학협력단
Abstract: PURPOSE: An apparatus for priority based message scheduling for a multi agent system and a method thereof are provided to quickly adapt to a user preference and a system environment change by scheduling message priority based on the importance of message workload. CONSTITUTION: An agent platform device including functions for managing an internal agent manages priority information of agents. All ACL(Agent Communication Languages) messages pass through the agent platform device. The agent platform device transfers the ACL messages to the agents which use the registered preference information. A main container(114) manages the agent platform apparatus. An agent descriptor(111) manages the agents information.
-
-
-
-
-
-
-
-
-
Search Results
120
records
(took 0.022 seconds)
