ReEvo: Large Language Models as Hyper-Heuristics with Reflective Evolution
About
The omnipresence of NP-hard combinatorial optimization problems (COPs) compels domain experts to engage in trial-and-error heuristic design. The long-standing endeavor of design automation has gained new momentum with the rise of large language models (LLMs). This paper introduces Language Hyper-Heuristics (LHHs), an emerging variant of Hyper-Heuristics that leverages LLMs for heuristic generation, featuring minimal manual intervention and open-ended heuristic spaces. To empower LHHs, we present Reflective Evolution (ReEvo), a novel integration of evolutionary search for efficiently exploring the heuristic space, and LLM reflections to provide verbal gradients within the space. Across five heterogeneous algorithmic types, six different COPs, and both white-box and black-box views of COPs, ReEvo yields state-of-the-art and competitive meta-heuristics, evolutionary algorithms, heuristics, and neural solvers, while being more sample-efficient than prior LHHs.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Traveling Salesman Problem | TSPLIB (test) | -- | 115 | |
| Capacitated Vehicle Routing Problem | CVRPLib Set X | Average Optimality Gap26.5 | 111 | |
| Online Bin Packing | Weibull distribution | Gap (%)0.17 | 63 | |
| Traveling Salesman Problem | TSP50 | Optimality Gap0.00e+0 | 58 | |
| Traveling Salesman Problem | TSP-100 | -- | 53 | |
| Capacitated Vehicle Routing Problem | CVRP N=100 | Objective Value16.293 | 50 | |
| Automated Heuristic Discovery | AHD Individual (Instance-wise) | Average Tardiness4.01e+3 | 28 | |
| Traveling Salesman Problem | TSP-200 | Optimality Gap0.216 | 28 | |
| Capacitated Vehicle Routing Problem | CVRP N=20 10,000 instances (test) | Objective Value5.619 | 26 | |
| Traveling Salesman Problem | TSP N=200 | Cost Gap2.56 | 24 |