Improved genetic identification algorithms on moisture absorption and desorption on building materials surface

The discrete Z transfer function model of moisture absorption and desorption on building materials surface is given. The identification mathematical model describes the dynamic moisture performance of building envelopes by materials surface´s air relative humidity input and materials moisture capacity output. An improved genetic algorithm for identifying the coefficients, order and time delay of moisture mode is developed in floating point representations. The population mutates before crossover in the computing process of the improved genetic algorithm, which is benefit to the convergence. To solve the slow convergence during the later evolution, the micro mutation algorithm is adopted. Through the improved genetic algorithms identification method, there get single input-single output difference equations and Z transfer function of moisture model. The results show the model and identification method in the improved genetic algorithm are valid and practicable.