Abstract: The needs of control system of reentrant manufacturing system are analyzed and the double-feedback-based collaborative control strategy for reentrant manufacturing system is proposed. To decrease the complexity of control system for manufacturing system, a layered & collaborative control mechanism is addressed, which includes the system layer coordination mechanism, machine group / machine layer one, and rescheduling one. Based on principal of mixed control, a multi-agent-based architecture of autonomous control system which integrates central coordination and local one is proposed so as to make the requirement of non-linear dynamic control of reentrant manufacturing system, it has the main characteristics including distribution, intelligence, robustness, openness, real-time, etc. To implement the proposed architecture of col-laborative control system, an HLA/CORBA-based multi-lever software framework is designed, which includes the base level, distributed object level, agent level and application level. The experiment results show the collaborative control mechanism above proposed takes precedence of the dynamic scheduling in machine groups and the scheduling of heuristic rule in the daily throughput of wafers and the utilization rate of machine groups.

Key words: Control mechanism, Collaboration, Manufacturing system, finite element analysis, reverse engineering, Complex components