OptiCom improves optimization by dynamically combining language model tools
OptiCom : A Unified Framework for State-Conditioned Composition in LLM-Driven Optimization
Artificial Intelligence
Summary
Solving hard problems often means trying many methods and picking the best ones as conditions change. The authors show that the success of these methods depends on the current situation. They created OptiCom, a system that uses a big language model to smartly combine different tools during the problem-solving process. Tests on many benchmarks show OptiCom works better than other approaches by adjusting both immediate steps and overall strategies as it learns.
What this means in practice
- •For software optimization teams: Integrate adaptive control of multiple optimization methods using OptiCom to improve solution quality under changing compute budgets and problem states.
- •For ai product developers: Build smarter AI assistants that optimize tasks by dynamically selecting and tuning language model-based search strategies during execution.$Commercial implications: Enables commercial AI tools that adapt optimization strategies for better performance and efficiency across diverse tasks.
Authors
Chenxing Wei, Sichen Liu, Lizhao Liu, Ningyuan Sun, Chen Bingzhou, Ying He, Bo Jiang, Fei Yu, Yao Shu
Abstract
Large language models (LLMs) are increasingly deployed to solve complex scientific and practical problems via iterative optimization. However, dynamically coordinating diverse search mechanisms as candidate quality, failure modes, and resource budgets evolve remains a critical open challenge. Targeted empirical diagnostics reveal that mechanism effectiveness is highly state-dependent. Motivated by this, we analyze how individual decisions drive final outcomes, decomposing the expected terminal improvement under a shared budget into cumulative decision opportunities minus cumulative selection losses. Guided by this opportunity-loss theoretical foundation, we propose OptiCom, a unified framework that represents LLM-driven optimizers within a shared configuration space: C=(A,Q,O,E,M,S), corresponding to artifact, query, operator, evaluation, memory, and strategy. Operating within this space, a fast LLM-based Optimization Controller dynamically composes immediate mechanisms through structured Action Packages, while a slower Strategy Adapter refines long-term selection preferences, operator weights, and templates based on accumulated trajectory feedback. Comprehensive evaluations across 32 benchmark groups demonstrate the superiority of framework: OptiCom achieves an average Max-score rank of 1.72 among 14 evaluated configurations, securing the top score in 23 groups. Ultimately, these results highlight the broad applicability and high extensibility of OptiCom as a general-purpose paradigm for robust LLM test-time scaling.