HyEvo: Self-Evolving Hybrid Agentic Workflows for Efficient Reasoning

HyEvo introduces an automated workflow generation framework that combines probabilistic LLM nodes for semantic reasoning with deterministic code nodes for rule-based execution. Using an LLM-driven multi-island evolutionary strategy, it iteratively refines workflow topology and logic through execution feedback. Experiments demonstrate significant improvements over existing methods, reducing inference costs by 19× and execution latency by 16× while maintaining superior performance on reasoning...