About the Journal
Aims and Scope. Journal of Reinforcement Learning and Autonomous Decision Systems provides a focused venue for research on policy learning, value methods, model-based RL, offline RL, safe exploration. The journal prioritizes technically substantive, internationally relevant work in which the principal novelty lies inside the stated computing scope rather than only in the application domain.
Core topics in scope include:
- policy learning
- value methods
- model-based RL
- offline RL
- safe exploration
- multi-step decision evaluation
- reinforcement learning
- sequential decision making
- autonomous agents
- control
Evidence and methodological expectations. AI studies should use strong contemporary baselines, clearly separated training/validation/test data, leakage controls, ablation or component analysis where relevant, uncertainty or error analysis, and external, cross-dataset, or cross-domain validation when feasible. Data and model provenance must be transparent.
Normally outside scope. Single-dataset demonstrations, accuracy-only comparisons, unvalidated model stacking, prompt demonstrations without systematic evaluation, and applications of standard models with no methodological or generalizable scientific contribution are normally outside scope.
Research integrity and reproducibility. Authors should disclose datasets, software, model or system configurations, experimental protocols, statistical procedures, ethical approvals where applicable, competing interests, funding, and any material use of generative AI. Data and code should be shared when legally and ethically possible, or the restriction must be explained.
