| 摘要 |
PROBLEM TO BE SOLVED: To shorten a prediction computation time and to execute stable control, in model prediction control. SOLUTION: A candidate of the optimum control input is searched by an optimum control input candidate searching part 31, a controlled variable is predicted using a dynamic characteristic model of a control object 20 by a model prediction part 33, and a deviation between the predicted controlled variable and a target value is found to be evaluated by an evaluation part 35, so as to determine the optimum control input. A searching range is set using the dynamic characteristic model of the control object by an optimum control input searching range setting part 4, in advance to the search of the optimum control input candidate, and useless search is thereby precluded from being searched. When starting this time prediction control, the last optimum control input is determined as the optimum control input in this time to shorten further the prediction requisition time, in the case where a deviation between the controlled variable predicted using the last optimum control input as the candidate of the optimum control input in this time and the target controlled variable is within a prescribed range. COPYRIGHT: (C)2006,JPO&NCIPI
|
