Certified Improvements to Best-Known Thomson-Problem Configurations with a Search Pipeline Built with Large Language Model Agents
Organizations: Independent Researcher · Georgia Institute of Technology · Nanjing University
Abstract
The Thomson problem asks for N unit charges on a sphere with minimum Coulomb energy; for large N the best configurations known are curated in the Cambridge Cluster Database, whose largest-N entries were set in October 2026. We report configurations for N = 912, 1382 and 2502 with energies 0.0036186, 0.0109730 and 0.0614033 below the database values, certified by exact integer arithmetic. All three are numerically strict local minima modulo rotations and keep the defect inventory and scar compositions of the database configurations; the database value for N = 912 had been reached by two different methods. Our search runs unconstrained replica-exchange basin hopping from the database configurations, with a shared pool of minima, recombination of improved regions after relabelling by geometry, and gradient-only polishing; replica-exchange basin hopping alone found the N = 912 and 1382 configurations, the former in one of five runs. We built and ran it with large language model agents (Claude, through a chat interface and the Claude Code agent), and we report every search, including those that failed. Two findings matter for others who search the database. Polishing alone lowers two entries by more than the threshold at which our pipeline reports a candidate, which produced two spurious records until candidates were compared with the polished database energies (the polished baseline). A local-relaxation speed-up passed its exactness test yet found nothing in a controlled comparison in which every arm with full relaxation went below the database value within 15 minutes. We distil three practices for computational discovery with large language model agents: calibrate the search code and verify candidates with code separate from the search, compare candidates with the polished baseline, and test design changes in controlled comparisons with pass criteria fixed in advance.
Figures & tables
| Role | Responsibilities |
| Authors | choice of problem and sizes; time boxes and stop rules; decisions to continue, stop, or change direction; requirement that every result pass verification by separate code; accepting or rejecting review comments; decisions to release; communication with the CCD maintainers |
| Chat assistant | literature search and reading; experiment design; the search code of every stage (energy kernel, basin hopping, pool and combiner, scheduler and local relaxation) and the verification and release scripts, delivered as packages with instructions for the agent; reading returned data and proposing decisions; drafting reports |
| Agent | deployment and calibration; launching and monitoring long runs; debugging and patches, including the combiner’s relabelling fix; verification runs; status reports; executing approved releases |
| Reviewing AI system | review of the plan (nine issues); a short independent check run at |
| Energy | Certified bound | Decrease ( ) | 5-/7-fold points | |||||
| 912 | CCD | 33 / 21 | ||||||
| new | 33 / 21 | |||||||
| 1382 | CCD | 48 / 36 | ||||||
| new | 48 / 36 | |||||||
| 2502 | CCD | 48 / 36 | ||||||
| new | 48 / 36 |
| Arm | first (wall-clock min) | best after 2 h |
| full | 4.3 | |
| full+ET | 1.5 | |
| hybrid | 13.8 | |
| local | – |
| Error | Origin | Detected by | Consequence and fix |
| Aggressive floating-point compiler optimisation removed the compensation terms of the summation routine (absolute energy error at ) | chat assistant | calibration against the official verifier’s scores of three published configurations, before any search | optimisation disabled for that routine only; search kernels unaffected |
| Duplicate detection by rounding energies to , below their noise, in a genetic-algorithm variant at | chat assistant | of offspring identical to the start, none accepted | tolerance band of ; run restarted |
| The stage-2 combiner compared raw point labels, which the walk permutes (aligned fraction ), so the best configurations were never combined; the stage-1 superposition script had relabelled points by geometry | chat assistant | combiner tried one pair in two hours | relabelling by geometry ( Algorithm 1 ) |
| The stage-2 status report first stated candidate pairs and superpositions, ignoring the combiner’s limit of ten configurations | AI systems | checking the report against the run logs | corrected to pairs and superpositions |
| Interpretation that the decreases came from independent local rearrangements | AI systems | transplanting each change separately ( Section 5 ) | claim withdrawn |
| Hypothesis that “lower symmetry is better”, inferred from a sequence of records at | authors | reviewing AI system: and are not nested | claim withdrawn |
Appendix figures & tables5 assets
Supplementary material from the paper’s appendix.
Appendix
| Component | Setting |
| Length scale | |
| Replicas, temperatures | 3; , , (geometric) |
| Moves (stages 2 and 3) | displacement of neighbours, , probability 0.55; rotation of neighbours by , probability 0.35; displacement of neighbours, , probability 0.10 |
| Moves (stage 1) | first two move types with probabilities 0.65 and 0.35 |
| Relaxation | L-BFGS (SciPy implementation, no bounds), memory 30, gradient tolerance , relative function tolerance , at most 30,000 iterations, on unnormalised coordinates |
| Polishing | step of Section 4 (the first Barzilai–Borwein step), capped at and halved when ; stop at (stage 1 polishing: the lowest of three L-BFGS minimisations to gradient tolerance , two of them after Gaussian perturbations of standard deviation ; released configuration: alternated with the second Barzilai–Borwein step ) |
| Stage | Processes h | Relaxations | Best | Notes | |
| Preliminary | 960 | – | five strategies; about core-hours used | ||
| 1 (screening) | 1382 | 25,839 | both runs below the CCD value, the first after 1.5 min | ||
| 2000 | 13,164 | – | lowest distinct minimum | ||
| 2502 | 8,796 | both runs below the CCD value, the first after 1.2 min | |||
| 3552 | 4,785 | – | lowest distinct minimum | ||
| 2502 | – | – | symmetry-constrained ( ); best |
| 1382 | 2000 | 2192 | 2502 | 3182 | 3552 | 4352 | 5792 | 5872 | 9911 | 9913 | |
| Core-hours | 7.5 | 7.9 | 7.5 | 7.5 | 6.9 | 7.0 | 7.0 | 8.0 | 7.0 | 7.0 | 6.5 |
| Relaxations ( ) | 42.1 | 31.9 | 26.8 | 24.5 | 17.8 | 16.2 | 13.2 | 11.0 | 9.2 | 5.5 | 5.2 |
| Relaxation | full | local | local | local | local | local | local | local | local | local | local |
| Origin of CCD value | CCD value | Relaxations | Runs below CCD value | Best | |
| 900 | October 2026 method | 116,226 | 0 | ||
| 912 | earlier, reproduced | 117,852 | 1 | ||
| 920 | earlier, reproduced | 113,817 | 0 | ||
| 932 | earlier, reproduced | 109,965 | 0 |
Metaheuristic Optimization
Metaheuristic optimization develops flexible search strategies for finding strong solutions to difficult problems when exact methods are too costly.
See all 123 papersLLM-Guided Program Evolution for Circle Packing: Breaking 10 Packomania Records for $28
Scientific discovery as meta-optimization: a combinatorial optimization case study
state space'' of theories and experiments, and an evaluation criterion based on quality, novelty, and validity. Large language models (LLMs) have enabled automated exploration of this state space, but we argue that simultaneous modification of the evaluation criteria is equally important. Here, we propose formalizing research as meta-optimization, where the optimization objective itself is also being optimized. Our key contribution is consensus objective aggregation,'' where a portfolio of LLM-generated objective functions is combined via correlation-weighted voting, allowing the evaluation criterion to evolve alongside the solutions. We apply this framework to algorithm discovery for planted 3-SAT problems based on digital memcomputing machines. The leading solver from the consensus search reduces median integration steps by more than relative to a tuned baseline at 2,560 variables. Objective ablations show that a well-chosen fixed expert objective can also discover strong solvers, while consensus supplies a stable ranking that mitigates reward hacking behaviors. By bringing objective engineering into the discovery loop, this framework offers a practical route toward more adaptable automated research.