Search papers, labs, and topics across Lattice.
The paper introduces OR-Space, a new benchmark designed to evaluate LLM agents in realistic industrial operations research (OR) workflows, moving beyond single-shot problem solving. OR-Space features persistent multi-artifact workspaces and multi-stage task lifecycles, including model construction, revision, and grounded explanation. Experiments using OR-Space will allow researchers to study the reliability and failure modes of LLM agents in complex, real-world OR scenarios.
LLMs that ace one-shot optimization problems often stumble when faced with the messy reality of iterative model building, revision, and explanation in industrial settings.
Large language model (LLM) agents are increasingly used to assist with operations research (OR) modeling, yet existing OR-oriented benchmarks often reduce evaluation to one-shot translation from a self-contained problem statement into a mathematical formulation or solver program. Such settings abstract away two characteristics of real industrial OR workflows: persistent multi-artifact workspaces and multi-stage task lifecycles. We introduce OR-Space, a full-lifecycle workspace benchmark for evaluating industrial optimization agents across model construction, model revision, and grounded explanation. Each instance is an executable workspace containing business documents, structured data, optional code artifacts, solver outputs, and task-specific evaluators distributed across interdependent files. OR-Space defines three task modes: Build, where agents construct solver-ready optimization models from heterogeneous artifacts; Revise, where agents modify existing models under changing requirements or solver feedback while preserving valid prior logic; and Explain, where agents answer grounded questions about solutions, constraints, and business implications using evidence spread across workspace artifacts. By combining persistent workspaces with lifecycle-oriented tasks, OR-Space evaluates whether agents can perform reliable optimization work beyond end-to-end text generation. We describe the benchmark design, evaluation protocol, and quality-control pipeline, and position OR-Space as a benchmark for studying the reliability, failure modes, and practical readiness of LLM agents in industrial OR workflows.
