Automatic generating numerical control rule using genetic-based with multiple critical evaluation

A method is proposed to automatically extract numerical control rules from the sensor data without the help of experts by means of a genetic algorithms (GA) based on multiplet critical evaluation to meeting different criteria. Every generated numerical rule is accumulated in a control table called a numerical rule-based controller. The numerical control table can be stored into the controller of embedded control system to construct a numerical rule-based embedded controller to meet real-time processing in the industry. The combination of multiple critics applies on the controller as fitness function of genetic learning in order to generate numerical rules to achieve multi-objective genetic process. Moreover, this paper apply an experimental design method which add a 'King strategy' to crossover operator of the standard GA in order to reduce the blindness of GA search processes and raise the convergence speed. An illustrative experiment is successfully made on the computer simulation. The experimental results reveal that the proposed approach is more efficient and more effective than the single objective.