Abstract
The pervasiveness of micro processor units (MPUs) in mostly controlled system coupled with recent rapid advances in MPU technology have created both benefits and problems. One of the problems is that it takes more man-hours to develop a program of a automatically controlled system. Our research group has developed and offered the method to support to design such a control program, thus requiring fewer man-hours, using evolutionary computation such as genetic algorithm (GA) and genetic programming (GP). However, there is no clear policy how to select the gene expression and how to design a fitness function in order to apply evolutionary computation to air conditioning system. In this paper, the method of selecting fitness function when the evolutionary optimization is applied to air conditioning is described.
