Timing control points via simulation for production systems under random disturbances

The paper deals with the problem of production control with disturbances. T he problem is applied to semi-automated production systems where the advanc ement of the process cannot be measured continuously but only in preset con trol (inspection) points. It is assumed that the production process can be simulated within a time interval of arbitrary size. Given the target amount needed, the due date, the routine control point, the actual accumulated pr oduction observed at that point, and the preset confidence probability (cha nce constraint) of meeting the deadline on time, the decision-maker has to determine the next control point. On-line control has to be carried out as rarely as possible but without missing the moment when the tendency to devi ate from the planned trajectory may endanger meeting the due date on time u nder a chance constraint. Determining the next control point is carried out via simulation with a constant time step. At each intermediate check-point , decision-making based on sequential statistical analysis has to be undert aken, either. (a) to proceed further and to examine the next check-point; or (b) to determine the check-point under consideration as the last moment bef ore the production process deviates from its target subject to the chance c onstraint. Thus, the next routine control point is determined. (C) 2001 IMA CS. Published by Elsevier Science B.V. All rights reserved.