PARAMETER IDENTIFYING SYSTEM OF LARGE SCALE SYSTEM MODEL

摘要

PURPOSE:To execute practical identification by using a step that estimates the probability of an assumed value of unknown parameter basing on transcendental information together with a step that estimates basing on observed information. CONSTITUTION:A function that estimates the probability of a concrete value on each parameter is set as transcendental information. A model variable value is derived by assuming the initial value of a parameter, and the degree of dissidence between the obtained the mudel variable and observed quantity is calculated, and each parameter is evaluated from the viewpoint whether it is effective to correct the dissidence or not. A parameter which is evaluated as low in probability and effective to correct dissidence between the model variable and observed quantity is selected as a parameter for regulating. By grasping the regulting parameter in group unit such as ''one to be largely increasd'' and ''one to be little increased'', the scale of a problem of optimizing is reduced.