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
    • Quadratic programming solving apparatus
    • 四分之一编程解决方案
    • JP2006018657A
    • 2006-01-19
    • JP2004196842
    • 2004-07-02
    • Mitsubishi Electric Corp三菱電機株式会社
    • NAKAMURA SHIZUKA
    • G06F19/00G06Q10/04
    • PROBLEM TO BE SOLVED: To obtain a quadratic programming solving apparatus solving at a high speed a quadratic programming problem. SOLUTION: The quadratic programming solving apparatus comprises: an input means setting the equation of the quadratic programming problem including a function inequality restriction; a conversion means converting the equation into the function equality restriction and a new upper and lower limit restriction for a variable by introducing a slack variable; a correction equation generation means generating a correction equation for obtaining the correction amount of the variable to close to the optimum solution for the quadratic programming problem; a contraction correction equation generation means generating a variable assumed to be on the boundary of the upper and lower limit restriction in the optimum solution from the correction equation and a low-dimensional contraction correction equation obtained by eliminating Lagrangian multiplier for the variable; and a correction amount calculation means obtaining the correction amount of the variable to close to the optimum solution by solving the contraction correction equation. The use of the contraction correction equation obtains an apparatus achieving operation processing-amount reduced high-speed processing in obtaining the optimum solution by repeatedly correcting the variable. COPYRIGHT: (C)2006,JPO&NCIPI
    • 要解决的问题:为了获得二次规划解决装置,高速求解二次规划问题。 二次规划求解装置包括:输入装置,设定包括函数不等式限制的二次规划问题的方程; 转换装置,通过引入松弛变量将方程转换为函数等式限制和变量的新的上限和下限限制; 产生用于获得所述变量的校正量以接近所述二次规划问题的最优解的校正方程的校正方程生成装置; 收缩校正方程式生成装置,生成假定为来自校正方程式的最优解的上限和下限边界的变量,以及通过消除该变量的拉格朗日乘数获得的低维收缩校正方程; 以及校正量计算装置,通过求解收缩校正方程来获得接近于最优解的变量的校正量。 使用收缩校正方程获得通过重复校正变量来获得最佳解的实现操作处理量减少的高速处理的装置。 版权所有(C)2006,JPO&NCIPI
    • 44. 发明专利
    • Retrieval system for state space
    • JP2004070662A
    • 2004-03-04
    • JP2002229078
    • 2002-08-06
    • Mitsubishi Electric Corp三菱電機株式会社
    • NAKAMURA SHIZUKAITAYA NOBUHIKOSHIMIZU TSUNEOKOSHIO MASANOBUMORI KAZUMI
    • G06N5/04G05B13/02
    • PROBLEM TO BE SOLVED: To solve the problem that since the retrieval system of a state space is configured to evaluate all states defined so as to be transitable from a "state under consideration", and to repeat the transition of the "state under consideration" to the state having the minimum evaluation function value among those states to discover the optimal solution, the retrieving efficiency is not satisfactory, and the large number of times of retrieval is required. SOLUTION: In a process for evaluating all states defined so as to be transitable from a "state under consideration", the high order M pieces of states whose evaluation function values are satisfactory are extracted (a processing block 5), and some of the changes of the M kinds of states generated due to the transition from the "state under consideration" to the M pieces of states are simultaneously generated with respect to the "state under consideration", and the state having the optimal evaluation function value is extracted (processing block 8), and the transition of the "state under consideration" is carried out (a processing block 9). Thus, it is possible to carry out the transition of the "state under consideration" with a distance 1 or more, and to increase the retrieval efficiency of the state space. COPYRIGHT: (C)2004,JPO
    • 45. 发明专利
    • State estimation system
    • 状态估计系统
    • JP2002373159A
    • 2002-12-26
    • JP2001179732
    • 2001-06-14
    • Mitsubishi Electric Corp三菱電機株式会社
    • NAKAMURA SHIZUKA
    • G06F17/17G06F17/10
    • PROBLEM TO BE SOLVED: To provide a state estimation system capable of performing rigid state estimation regardless of a way to apply the initial condition of weight to an inputted measured value which is actually measured in performing the state estimation of a system.
      SOLUTION: The weighting sum of squares Σwi(zi-hi(x))
      2 of a difference between a measured value zi which is actually measured and an estimated value hi(x) of the measured value which is calculated from the estimated value of state variables x is defined as an evaluation function, and the initial values of the state variables x and the weight wi are respectively set, and the evaluation function is minimized by using the state variables x and the weight xi as variables under a constraint condition that the product of the weight xi is fixed. Thus, the estimated value of the optimal state variables x and optimal weight w can be obtained.
      COPYRIGHT: (C)2003,JPO
    • 要解决的问题:提供一种能够执行刚性状态估计的状态估计系统,而不管如何对在执行系统的状态估计时实际测量的输入测量值应用初始重量条件。 解决方案:实际测量的测量值zi与从估计值计算的测量值的估计值hi(x)之间的差的平方Σwi(zi-hi(x))<2> 状态变量x被定义为评估函数,并且分别设置状态变量x和权重wi的初始值,并且通过使用状态变量x和权重xi作为约束条件下的变量来最小化评估函数 重量xi的产品是固定的。 因此,可以获得最优状态变量x和最佳权重w的估计值。
    • 49. 发明专利
    • Calculating method of inverted matrix
    • 计算矩阵的计算方法
    • JPS6111883A
    • 1986-01-20
    • JP13189684
    • 1984-06-28
    • Mitsubishi Electric Corp
    • NAKAMURA SHIZUKA
    • G06F17/16G06F15/78
    • G06F15/8053
    • PURPOSE:To attain rapid processing and to obtain various circuit constitutions in accordance with similarity or a processing speed by finding out a value after the change of an inverted matrix which is caused by the change of a matrix through product circuits, an integrating circuit and a subtractor. CONSTITUTION:The value of CX(s) is calculated by the 1st product circuit 1a and its integrated result is outputted to a point alpha. Then, the CX(s) and X(s) are integrated by the 2nd product circuit 1b and the integrated result is outputted to a point beta. Consequently, the value of integral X(s)CX(s)ds is found out at a point gamma by integrating the said output by an integrating circuit 2 and the value of (A+SC) , i.e. the value X(s) of the inverted matrix corresponding to the matrix obtained after changing its initial value, is calculated by subtracting the value of integral X(s)CX(s)ds at the point delta from the inverted matrix A corresponding to the initial value of the matrix by the subtractor 3a.
    • 目的:通过在通过乘积电路的矩阵变化引起的反转矩阵变化之后,找出相应于相似度或处理速度的各种电路结构,获得快速处理,积分电路和 减法器 构成:CX(s)的值由第一乘积电路1a计算,其积分结果输出到α点。 然后,CX(s)和X(s)由第二乘积电路1b积分,积分结果被输出到点β。 因此,通过积分电路2对所述输出和(A + SC)<1>的值进行积分,即在点γ处求出积分X(s)CX(s)ds的值,即值X 通过从对应于对应于(1)的倒矩阵A -1的点delta减去积分X(s)CX(s)ds的积分X(s)ds来计算与改变其初始值之后获得的矩阵相对应的反转矩阵的一个或多个 减法器3a的矩阵的初始值。