From Multi-Agent Reinforcement Learning to Agentic AI: A Comprehensive Literature Review of Algorithmic Advances and Decision-Analytic Implications (2020-2025)

Integration of Multi-agent Reinforcement Learning (MARL) and Large Language Models (LLMs) in Agentic AI Systems: A Narrative Review

AGTA: Topology-Aware Sequential Decision-Making in Multi-Agent Reinforcement Learning

Action Generation with Topology Awareness (AGTA), a topology-aware sequential decision-making framework in MARL that integrates inter-agent correlation modeling with topology-guided decision-order optimization, and outperforms the state-of-the-art counterparts.

Reward-Driven LLM Agent Workflows: Synthesizing POMDP Routing and Self-Correction for Autonomous Decision-Making

This research bridges theoretical foundations of reinforcement learning and graph-based memory with autonomous agent workflows, and offers a practical, scalable reference framework for developing artificial intelligence technologies in complex, multi-step autonomous systems.

Reinforcement Learning: From Algorithms To Foundation Models

Reinforcement learning (RL) provides a framework for sequential decision making under explicit objectives. In its classical form, RL studies how an agent should act to maximise long-term reward in a dynamic environment. In richer settings, the problem extends beyond a single agent and fixed environment: intelligent behavior may require strategic interaction, adaptation to uncertainty, and reasoning over high-dimensional worlds. This thesis studies RL from two perspectives: algorithms in games and RL in the era of foundation models. The first part focuses on multi-agent RL in games. It examines how incentives, policies, and equilibrium concepts interact in competitive and general-sum environments, spanning two-player zero-sum games, large-scale video games, and multi-player settings with general structure. These works investigate learning in multi-agent systems and the behavior of RL methods in interactive environments. The second part studies RL with generative and foundation models, motivated by the idea that prior knowledge can enrich sequential decision making. Pretrained generative models and learned world models serve as representation tools and structured priors for planning, control, and policy optimization. The thesis develops diffusion-based world models, investigates RL for efficient video generation, explores generative models as policy classes, and studies interactive video world models in which actions shape future observations. It also addresses long-horizon modeling through architectures with memory. Together, these contributions present a unified view of RL as objective-driven adaptation in complex sequential domains. From strategic games to generative world models, the thesis highlights how RL connects decision making, environment modeling, and emerging foundation-model capabilities, offering a broader perspective on the principles underlying intelligent behavior.

Large Language Model (LLM) as an Excellent Reinforcement Learning Researcher in both Single-Agent and Multi-Agent Scenarios

In the quantitative finance area, particularly in order execution, reinforcement learning (RL) has shown great promise due to its ability to interact with market environments based on real data. However, traditional RL methods suffer from slow research speed and rely on static market assumptions, which do not consider the impact of the agent's execution action on the environment. To address these, we propose a Self-Evolutional single-agent/multi-agent Reinforcement Learning (SE-RL) framework. The framework utilizes a Large Language Model (LLM) to design various RL algorithm modules, such as agent model design, reward function, profiling, communication, and state imagination, by leveraging the LLM generating module output or code. SE-RL could continuously improve the accuracy of LLM-generated RL algorithms through a dual-enhancement kit at both high-level (prompt refinement) and low-level (parameter fine-tuning). Additionally, we use a multi-agent system to simulate dynamic financial markets, accounting for the impact of order executions on market dynamics. To further enhance training in such a dynamic market, we develop a hybrid environment training method that could rebalance each environment's loss weight. Comprehensive experiments on 200 realistic stock datasets demonstrate that our proposed framework outperforms current state-of-the-art baselines. Project page: https://kdd2026-se-rl.github.io/.

Agent-Based Modeling and System Dynamics Integrated with AI and Analytical Methods: A Structured Review of Hybrid Approaches, Applications, and Future Directions for Decision Making in Complex Systems (2021–2026)

The complexity of socio-economic, ecological, and public-health systems demands simulation frameworks capturing both macro dynamics and micro agent heterogeneity. Agent-based modeling (ABM) and system dynamics (SD), increasingly coupled with AI and analytical methods, support decision making in complex systems, yet no systematic review covers 2021–2026. A two-round PRISMA 2020 search (September 2025; May 2026) in IEEE Xplore, Web of Science, and ProQuest identified 70 eligible papers across 14 thematic clusters; because the second search round closed in May 2026, papers published later in 2026 are necessarily under-represented relative to complete prior years. Data were extracted along eleven dimensions; inter-rater reliability was κ = 0.88–0.90. Output grew 133% between 2023 and 2025. C14 (Multi-Method Simulation) and C7 (Healthcare & Epidemiology) are the largest clusters; AnyLogic dominates as the only natively tri-paradigm platform (22.9%). Hybrid models consistently identify critical intervention thresholds invisible to mono-paradigm approaches. Only 44.3% of papers report formal structural validation and 88.6% withhold code. Nine papers (12.9%) integrate ML/AI, but none apply interpretability techniques (SHAP, LIME, ICE). ABM–SD hybridization is a maturing paradigm whose epistemic gain depends on closing validation, reproducibility, and interpretability gaps. Six gaps and five priority directions for 2026–2030 are identified, including a minimum validation protocol, a standardized repository, and a dedicated interpretability framework for ML+ABM–SD hybrids that preserves causal transparency for decision making in complex systems.