This article compares various evolutionary algorithms for knowledge discovery based on Compensatory Fuzzy Logic through fuzzy predicates. One of them, called EU-CFL-GP, is the result of the integration of two independent optimization algorithms (EK-CFL and EO-GSF) where EK-CFL searches high truth value predicates, and the EO-GSF optimizes membership functions by matching them to available data. For the EU-CFL-GP algorithm, an improvement implements through a mutate operator through constructive mutation. These algorithms are based on compensatory Fuzzy Logic, which allows the modeling of vagueness that is carried out through linguistic labels. It will enable expressions in a sophisticated way close to human thinking, solving the problem for decision-making examples in the analytics of business.