This package contains the datasets, experimental results and source code of the paper <em>Effects of the Random Forests Hyper-Parameters in Surrogate Models for Multi-Objective Combinatorial Optimization: A Case Study using MOEA/D-RFTS</em>. The following files are included: <ul> <li>File <em>datasets.zip</em> - contains the datasets used to train and test the Random Forest in an online learning process. Each dataset contains 5100 instances (decision vectors), all of them with the corresponding objective functions. There is a dataset for each benchmark problem: (i) the Binary Multi-Objective Knapsack Problem (BIN_MOKP); (ii) the Binary Multi-Objective Unconstrained Combinatorial Optimization Problem (BIN_MUCOP), and (iii) the Integer Multi-Objective Unconstrained Combinatorial Optimization Problem (INT_MUCOP). The file names are organized as follows: <problem name>_M<number of objectives>_<number of decision variables>. One example is: bin_mokp_M2_100.csv, indicating that this dataset is used for the BIN_MOKP problem, with M=2 objectives and 100 decision variables;</li> <li>File <em>experiments_results.zip</em> - contains the experimental results of the predictions and also the optimizations. The subfolder “prediction” contains Mean Absolute Error (MAE) results of each hyper-parameter combination. The file names are organized as follows: <problem name>_M<number of objectives>_<number of decision variables>_tunning_results.csv. One example is: bin_mokp_M2_100_tunning_results.csv, indicating that this file contains the results of the predictions when we applied the Random Forest to the BIN_MOKP problem, with M=2 objectives and 100 decision variables. The subfolder “optimization” contains the results of the optimization process for each algorithm, problem and dimensionality;</li> <li> File <em>source_code.zip</em> - contains the source code (in Python Programming Language) to reproduce the experiments. The source code contains: (i) the implementations of the algorithms MOEA/D, MOEA/D-NFTS and MOEA/D-RFTS; (ii) the test instances of each benchmark problem (BIN_MOKP, BIN_MUCOP and INT_MUCOP), number of objectives (2 and 3) and dimensionality (100,300 and 500). The instructions are in the README.txt file inside the package. </li> </ul>