A novel architecture for controlling and managing large-scale intelligent systems is presented, in which the different expert systems and numerical computation routines are coordinated by a metasystem. These expert systems and numerical routines may be written in different languages or programming tools, debugged and used separately. In this way, it is possible to easily add new programs and reduce the scope of rule search to enhance efficiency. Moreover, the optimal solution can be selected among the conflict results, and parallel processing becomes feasible in the integrated intelligent system. It is concluded that this architecture can serve as a universal configuration to develop high-performance intelligent systems for many complicated industrial applications in real-world domains.
