DAO-CITO: A Dual Aggregation-Oriented Method for Reinforcement Learning-based Class Integration Test Order Generation

Description This artifact accompanies the paper "DAO-CITO: Dual-Aggregation Reinforcement Learning for Class Integration Test Optimization". DAO-CITO is a dual-aggregation reinforcement learning approach that improves the learning efficiency of Class Integration Test Order (CITO) generation by jointly exploiting dependency-aware action aggregation and value-guided state aggregation while preserving the quality of the generated test orders. The artifact includes: the complete Python implementation of DAO-CITO; the processed benchmark datasets used in the experiments; the scripts for training and evaluation; the experimental results of DAO-CITO and all compared approaches; the statistical analysis results reported in the paper. The processed datasets are derived from publicly available Java benchmark projects and are included in the artifact. No additional preprocessing is required to reproduce the reported experiments. Detailed instructions for installation, execution, and reproduction are provided in the accompanying README file. This artifact is intended to support the reproducibility of the experimental results reported in the paper.