pattern based adaptive uniform recombination in XCS

recombination in the accuracy -based learning classifier system XCS which is introduced by wilson in 1995 has always been done by the means of one or two point recombination operators. This paper investigates and compares the performance of a recently intorduced recombination operator with these traditional ones in boolean functions in this paper we are trying to use a new crossover operator that is defined in our previous works this operator is called PAUX and is based on statistical information of populationin our previous experiments this operator showed significantly better performance than other traditional ones. in solving boolean problems there fore we try to use this new recombination method as the main recombination operator in XCS and to investigate its advantages.