Evolutionary algorithms such as Genetic Algorithm (GA), Non Dominated Sorting Genetic Algorithm-II (NSGA-II) have been widely recognized as effective approaches for optimization problems. Recent developments have introduced integration of learning based operators, notably the Innovized Progress (IP) operator to improve the convergence characteristics. In this research, we have implemented multiple evolutionary algorithms including Vector Evaluated Genetic Algorithm (VEGA), NSGA-I, NSGA-II and NSGA-II with IP operator. Additionally, we have developed and implemented NSGA-II IP variants by replacing the original Artificial Neural Network (ANN) model with CatBoost Regressor (CBR) and Light Gradient Boosting (LGB) models. All algorithms are evaluated on a set of standard benchmarked functions with performance measured using hypervolume, convergence, and spread metrics. Our results demonstrate that the newly introduced LGB-based variant achieves the highest hypervolume on ZDT2, better convergence on ZDT6, and superior spread on both ZDT6 and DTLZ4. Meanwhile, the CBR-based variant excels in maintaining a better spread on ZDT2. These findings indicate that the choice of ML integrated variant is significantly influenced by the specific characteristics of the optimization problem.