SDDL Neuro-Symbolic Framework Improves Scheduling Feasibility in Small LLMs

SDDL translates natural-language scheduling problems into solver-aligned representations, delegating search to deterministic compilers. On 300 multi-family scheduling instances, the framework improves independently verified feasibility for all tested resource-constrained models.