On-line job-shop scheduling of a manufacturing system based on a virtual supervisor concept

The controls for reconfigurable manufacturing systems have to be capable not only of identifying exceptions on-line, but also simultaneously developing on-line strategies for unpredictable customer orders or inaccurate estimate of processing times. This paper presents an approach for job-shop scheduling with uncertain arrival times. The approach exploits Virtual Supervisor (VS) concept, which has access to all system information during program execution and thus can readily monitor the overall system performance. The goal is to minimize expected part tardiness and earliness cost. A solution methodology based on a combined Lagrangian relaxation, VS-Patterns, Maxwell equations and temporal difference is developed to obtain a dual solution for on-line implementation.