From Static Templates to Dynamic Runtime Graphs: A Survey of Workflow Optimization for LLM Agents

ArXi:2603.22386v1 Announce Type: new Large language model (LLM)-based systems are becoming increasingly popular for solving tasks by constructing executable workflows that interleave LLM calls, information retrieval, tool use, code execution, memory updates, and verification. This survey reviews recent methods for designing and optimizing such workflows, which we treat as agentic computation graphs (ACGs