DyFlow: Dynamic Workflow Framework for Agentic Reasoning

2025-09-30

AI Summary

Existing LLM-based agent workflows often rely on manual design or are constrained to narrow tasks, exhibiting poor generalization, underutilization of intermediate feedback, and rigid execution logic. To address these limitations, we propose DyFlow—a dynamic, adaptive workflow framework that decouples planning from execution. DyFlow employs context-aware operator parameterization and LLM-driven subgoal decomposition and path re-planning to dynamically construct and adjust workflows in response to task requirements and real-time reasoning feedback. Its core innovations include a semantics-driven dynamic operator mechanism and an extensible workflow generation paradigm. Extensive experiments across diverse domains—including social reasoning, biomedical analysis, mathematical problem solving, and code generation—demonstrate that DyFlow significantly improves Pass@k performance. It exhibits strong cross-task generalization and robust deep-reasoning capabilities, outperforming static and manually engineered alternatives.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Multiagent Systems: Multiagent Planning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Agent systems based on large language models (LLMs) have shown great potential in complex reasoning tasks, but building efficient and generalizable workflows remains a major challenge. Most existing approaches rely on manually designed processes, which limits their adaptability across different tasks. While a few methods attempt automated workflow generation, they are often tied to specific datasets or query types and make limited use of intermediate feedback, reducing system robustness and reasoning depth. Moreover, their operations are typically predefined and inflexible. To address these limitations, we propose DyFlow, a dynamic workflow generation framework that adaptively constructs and adjusts reasoning procedures based on task requirements and real-time intermediate feedback, thereby enhancing cross-task generalization. DyFlow consists of two core components: a designer and an executor. The designer decomposes complex problems into a sequence of sub-goals defined by high-level objectives and dynamically plans the next steps based on intermediate outputs and feedback. These plans are then carried out by the executor, which executes each operation using dynamic operators with context-aware parameterization, enabling flexible and semantically grounded reasoning. We systematically evaluate DyFlow across diverse domains, including social reasoning, biomedical tasks, mathematical problem solving, and code generation. Results demonstrate that DyFlow significantly outperforms existing baselines, achieving substantial Pass@k improvements and exhibiting robust generalization across diverse domains. The code is publicly available at https://github.com/wyf23187/DyFlow.

Problem

• Dynamic workflow generation for agentic reasoning systems
• Adaptive reasoning procedures using real-time feedback
• Cross-task generalization in complex reasoning domains

Innovation

• Dynamic workflow generation with real-time feedback
• Designer decomposes problems into sub-goals dynamically
• Executor uses context-aware parameterization for operations