MultiHedge: Adaptive Coordination via Retrieval-Augmented Control
For decision-making systems under uncertainty, this work demonstrates that architectural design (retrieval) can be more effective than scaling, though the evaluation is limited to a specific domain and baselines.
MultiHedge uses retrieval-augmented LLM coordination to improve robustness in modular decision pipelines, showing in U.S. equities that memory-augmented retrieval yields greater stability than increasing model scale alone.
Decision-making under changing conditions remains a fundamental challenge in many real-world systems. Existing approaches often fail to generalize across shifting regimes and exhibit unstable behavior under uncertainty. This raises the research question: can retrieval-augmented LLM coordination improve the robustness of modular decision pipelines? We propose MultiHedge, a hybrid architecture where an LLM produces structured allocation decisions conditioned on retrieved historical precedents, and execution is grounded in canonical option strategies. In a controlled evaluation using U.S. equities, we compare MultiHedge to rule-based and learning-based baselines. The key result is that memory-augmented retrieval confers greater robustness and stability than increasing model scale alone. Our paper contributes a controlled computational study showing that memory and architectural design play a central role in robustness in modular decision systems.