Genetic Programming of a Microcontrolled Water Bath Plant

Typically, control system design leads to a higher-order non-linear function of the system's state variables. As a result, it is very hard to find a satisfactory mathematical solution. On the other hand, considering a microcontroller based implementation, another difficulty is to program it to carry out the desired control algorithm. This paper presents the application of linear genetic programming in the automatic synthesis of a microcontroller assembly program, which performs an optimized control of a water bath plant. The synthesis starts from the plant's mathematical modeling and supplies directly a assembly code for the microcontroller platform. When comparing the control performance of the synthesized program with that of a neuro-fuzzy based controller, the synthesized program proved to perform slightingly better.