会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 42. 发明专利
    • Tape storage device, data writing method, and program
    • 磁带存储设备,数据写入方法和程序
    • JP2012243353A
    • 2012-12-10
    • JP2011112885
    • 2011-05-19
    • Internatl Business Mach Corp インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation
    • ABE ATSUSHIHAGIWARA KATSUHIKO
    • G11B20/10G06F3/06
    • G06F3/0613G06F3/0659G06F3/0682G11B5/008G11B2220/90
    • PROBLEM TO BE SOLVED: To receive subsequent write data without a backhitch after a synchronization request.SOLUTION: A tape drive receives write requests and a synchronization request of a plurality of data pieces from a host device, performs a synchronization process (storing buffer data in a tape medium), and returns a completion of the synchronization request to the host device. The tape drive includes: a write control that transfers data from the buffer to the tape medium; and a command queuing function to asynchronously process a plurality of commands issued in succession from the host device. The write control is configured to (a) store data, transferred according to a first write request, in the buffer and receive a write request for a data set following a first synchronization request, and (b) return a completion of processing of the first synchronization request after receiving the write request by use of the queuing function. The write control stores, as a position of the synchronization request, a buffer position at which the last data is located at the time of receiving the synchronization request. The completion of the first synchronization request is returned after a subsequent second write request is received.
    • 要解决的问题:在同步请求之后接收后续写入数据而没有后退。 解决方案:磁带驱动器从主机装置接收多个数据的写请求和同步请求,执行同步处理(在磁带介质中存储缓冲数据),并将同步请求的完成返回给 主机设备。 磁带驱动器包括:将数据从缓冲器传送到磁带介质的写入控制; 以及命令排队功能,以异步处理从主机设备连续发布的多个命令。 写控制被配置为(a)将根据第一写请求传送的数据存储在缓冲器中,并且在第一同步请求之后接收对数据集的写请求,以及(b)返回第一 通过使用排队功能接收到写请求后的同步请求。 作为同步请求的位置,写入控制在接收到同步请求时存储最后数据所在的缓冲器位置。 在接收到后续的第二写请求之后返回第一同步请求的完成。 版权所有(C)2013,JPO&INPIT
    • 44. 发明专利
    • Information processing device, calculation method, program, and recording medium
    • 信息处理设备,计算方法,程序和记录介质
    • JP2012220819A
    • 2012-11-12
    • JP2011088093
    • 2011-04-12
    • Internatl Business Mach Corp インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation
    • IMAMICHI TAKASHIIDE TAKESHINUMATA HIDETOSHI
    • G02B5/02
    • G02B6/0065G02B6/004G02B6/0043G02B6/0061G02F1/133606G06T11/00
    • PROBLEM TO BE SOLVED: To generate a uniformly distributed discrete pattern.SOLUTION: An information processing device 100, for calculating a spatial arrangement of plural elements of a discrete pattern formed by the elements spatially discretely arranged, calculates, for each element, a density at initial positions respectively given to the individual elements on the basis of density distribution of the elements within a region where the elements of the discrete pattern are arranged, to set a diagram having a size in accordance with the density and representing a region in which the individual elements reject others and the moving range of the diagram at the initial positions of the individual elements, and minimizes a target function to give a penalty to a mutual collision of diagrams and protrusion of a diagram from the moving range, using the respective positions of the individual diagrams as decision variables to calculate the optimum solution of the respective positions of the individual diagrams and output the resultant optimum solution of the respective positions of the individual diagrams as the spatial positions of the elements.
    • 要解决的问题:产生均匀分布的离散模式。 解决方案:一种信息处理装置100,用于计算由空间上离散布置的元素形成的离散图案的多个元素的空间排列,对于每个元素,计算分别给予各个元素的初始位置处的密度 设置离散图案的元素的区域内的元素的密度分布的基础,以设置具有根据密度的尺寸的图并且表示各个元素拒绝其他区域的区域和图的移动范围 在单个元素的初始位置处,并且使目标函数最小化以给图表的相互碰撞和图形从移动范围突出的惩罚,使用各个图的各个位置作为决策变量来计算最优解 各个图的相应位置,并输出所得到的最佳解 单个图表的位置作为元素的空间位置。 版权所有(C)2013,JPO&INPIT
    • 50. 发明专利
    • Action determination method, program, and system
    • 行动决策方法,程序和系统
    • JP2012168767A
    • 2012-09-06
    • JP2011029660
    • 2011-02-15
    • Internatl Business Mach Corp インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation
    • OSOGAMI TAKAYUKIMICHAEL ONFAIT
    • G06Q40/04
    • G06N7/005G06Q40/00
    • PROBLEM TO BE SOLVED: To provide technique for selecting an action which approximately minimizes a repetitive risk index by using a Monte Carlo method in a Markov decision process of which scale cannot be precisely optimized.SOLUTION: Based on a Markov decision process, a queue of states (S,S,...,S) is sequentially generated by sampling. Then, in reverse order of the generated queue (S,...,S,S), values (provisional values of repetitive risk indexes) of respective states (V(S),...,V(S),V(S)) are updated. At this time, a value V(S) of each state Sis updated according to a value of a risk index of a probability variable (particularly calculated by using Value at Risk, an excess probability or a part of them) which is defined by a transition probability p(S|S) and a value of a transition destination V(S) of each of the destination states (S,S,...,S) which can be reached by one transition from each state S.
    • 要解决的问题:提供用于选择在马尔可夫决策过程中使用蒙特卡罗方法近似最小化重复风险指数的动作的技术,其中规模不能被精确地优化。

      解决方案:基于马尔可夫决策过程,状态队列(S 1 ,S 2 ,... ,S n )。 然后,按照生成队列的相反顺序(S n ,...,S 2 ,S n ),...,V(S 2 ),V(S 1 ))更新。 此时,根据风险指数的值更新各状态S i 的值V(S i ) 通过转移概率p(S i + 1 (j)¾S i )和转换目的地V的值(S i + 每个目的地状态的SB> (j))(S i + 1 1),S i + 1 (2),...,S i + 1 (m)),可以从每个状态S i 进行一次转换。 版权所有(C)2012,JPO&INPIT