发明名称 Computing method and computing apparatus
摘要 In a computing apparatus, a computing unit performs statistical analysis on data including the values of a plurality of input variables and the values of a plurality of output variables optimized by adjusting the values of the plurality of input variables, thereby generating the values of intermediate variables that indicate relative positional relationship among the values of each of the plurality of input variables and the plurality of output variables. The creation unit creates scatter diagrams representing the tendency of magnitude relationship among values of a variable, by using the generated values of the intermediate variables, for the respective input variables and output variables. The display unit displays the created scatter diagrams.
申请公布号 US9495776(B2) 申请公布日期 2016.11.15
申请号 US201213606991 申请日期 2012.09.07
申请人 FUJITSU LIMITED 发明人 Matsumoto Kazuhiro
分类号 G06T11/20;G06Q10/04;G06Q50/04;G06Q10/00;G06F19/00;G06F17/00 主分类号 G06T11/20
代理机构 Fujitsu Patent Center 代理人 Fujitsu Patent Center
主权项 1. A computing method comprising: calculating, by a processor, distances representing degrees of dissimilarity between different records included in a set of data, the records each being a combination of values of a plurality of input variables and values of a plurality of output variables produced by a computer simulation using the values of the input variables, wherein the values of the input variables are obtained as a result of multi-objective optimization performed with respect to the computer simulation; producing, by the processor, coordinates from the calculated distances between the records, the coordinates representing relative positional relationships among the records;creating, by the processor, scatter diagrams for the respective input variables and output variables, the scatter diagrams each plotting samples representing values of the input variables and output variables at the coordinates produced for the records corresponding thereto;displaying, by the processor, the created scatter diagrams;dividing, by the processor, a region of the scatter diagrams into subregions, and generating intersection coordinates that indicate positions of intersection points of the subregions in the scatter diagrams;calculating, by the processor, approximate values of the output variables by using approximate formulas for approximating the output variables, based on the generated intersection coordinates;selecting, by the processor, a subset of the generated intersection coordinates that matches better with a predetermined condition about the calculated approximate values than other generated intersection coordinates in terms of at least one of the output variables;calculating, by the processor, approximate values of the input variables by using approximate formulas for approximating values of the input variables, based on the selected subset of the generated intersection coordinates;generating, by the processor, values of the output variables for each of the intersection coordinates belonging to the selected subset, by executing the computer simulation using the calculated approximate values of the input variables; andcreating, by the processor, a scatter diagram in which samples representing the generated values of the output variables are plotted at the positions of the intersection points.
地址 Kawasaki JP