Abstract:

The simple genetic algorithm which uses a fixed crossover probability and mutation probability can not always meet the current needs of population, which affects the algorithm performance and efficiency. This paper uses the adaptive crossover probability and mutation probability, unifies the parallel technology and the genetic algorithm, and proposes the adaptive parallel genetic algorithm for Poisson curve settlement forecasting model optimization. The experimental results indicate that, this algorithm provides an effective method for the parameter estimation of Poisson curve settlement forecasting model.

Key words: adaptive, parallel genetic algorithm, Poisson curve, parameter estimation