Path planning of robot soccer based improved pseudo-parallel genetic algorithm

Through the analysis of existing pseudo-parallel genetic algorithm, proposing a pseudo-parallel genetic algorithm of the new dynamic sub-population, which changes the condition that the scale of sub-population is stationary in current existing information exchange model, the scale of sub-population will vary with the evolution. This algorithm can not only restrain premature convergence but also get the best global values and local values of multi-objective functions rapidly. Design the adaptive crossover operator according to the numbers of evolution generation. The crossover probability will be adjusted automatically according to the evolution, which accelerates the speed of the convergence. Through testing functions, the accuracy and superiority of this algorithm are proved. The simulation shows that the algorithm proposed is reliable and efficient for the path planning of robot soccer.