| 摘要 |
<p>A parallel distributed processing network (14) previously trained to simulate a process, is responsive to input activations A (42) to produce outputs Y (40). The activations are representative of the input parameters P (16) of the process (12) and the outputs are representative of the characteristics C (18) of the product of the process (12). A set of goal signals G (26) corresponding to the desired physical characteristics D (20) of the product are produced and used to generate a signal representative of the error E between the goal signals G (26) and the outputs Y (40) of the network (14). The values of the updated inputs AU (44) of the network (14) that are needed to drive the error E (38) to a predetermined minimum are determined, and the actual physical input parameters P (16) of the process changed to correspond to the updated input values AU (44) producing the minimum error. The present invention can be used in both a batch utilization situation and/or in an on-going process control situation.</p> |
