ACCORD rewrites combinatorial optimization solutions as step-by-step state updates that keep constraints visible, and this representation raises solution feasibility and lowers optimality gap versus list-of-list and prompting baselines.
Shifting bottleneck procedures for job shop scheduling
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ACCORD: Autoregressive Constraint-satisfying Generation for COmbinatorial Optimization with Routing and Dynamic attention
ACCORD rewrites combinatorial optimization solutions as step-by-step state updates that keep constraints visible, and this representation raises solution feasibility and lowers optimality gap versus list-of-list and prompting baselines.