feat(StructuralOptimization): add composite laminate stacking benchmark

[Mingqwqqaq]

This PR adds StructuralOptimization/CompositeLaminateStacking, a discrete composite-laminate design benchmark for Frontier-Engineering.

The task evaluates whether an agent can optimize balanced, symmetric 48-ply laminates across multiple plate geometries and compression load ratios. Candidate designs are scored using the governing reserve factor from maximum-strain failure and simply supported plate buckling analyses.

The benchmark is derived from the MIT-licensed research dataset:

• Dataset for Beyond Double-Double Theory: n-Directional Stacking Sequence Optimisation in Composite Laminates
• DOI: 10.5281/zenodo.15864525

Key Features

• Ten deterministic cases combining five aspect ratios with uniaxial and biaxial compression.
• Twelve integer angle decisions per case, giving 120 discrete design variables.
• Automatic expansion into balanced, symmetric 48-ply stacking sequences.
• Independent NumPy-only implementation of:
  • classical lamination theory and A/B/D matrices;
  • maximum-strain failure load factors;
  • double-sine Ritz buckling analysis;
  • symmetric generalized eigenvalue solution.
• Robust scoring based on both mean and twentieth-percentile case performance.
• Actionable per-case validation feedback for invalid or incomplete outputs.
• Candidate import and execution in a bounded worker process to reduce state leakage and evaluator tampering.
• English and Chinese documentation, provenance notes, baseline results, regression tests, and unified evaluation metadata.

Quality Assurance

• Direct baseline evaluation:
  • 10/10 feasible cases
  • combined_score = 50.0
• Unified OpenEvolve evaluation:
  • benchmark_returncode = 0
  • valid = 1
  • feasible_cases = 10
  • combined_score = 50.0
• Unit tests:
  • 13/13 tests passed
• Regression coverage includes:
  • balanced and symmetric layup construction;
  • negligible laminate coupling matrix;
  • upstream reference cross-checks;
  • invalid schema and angle handling;
  • syntax and timeout failure handling;
  • candidate process isolation;
  • monotonic score normalization;
  • measurable optimization headroom above the baseline.
• git diff --check passes.

Reproduction

From the task directory:

python verification/evaluator.py scripts/init.py
python -m unittest discover -s verification -p "test_*.py" -v

From the repository root:

python -m frontier_eval \
  task=unified \
  task.benchmark=StructuralOptimization/CompositeLaminateStacking \
  algorithm=openevolve \
  algorithm.iterations=0

[github-actions]

🤖 AI Code Review (gemini-3-flash-preview)

🇬🇧 English Analysis

1. Executive Summary

• Core Purpose: This PR introduces a new engineering benchmark task named CompositeLaminateStacking under the StructuralOptimization category. The task requires an agent to optimize the stacking sequence of a 48-ply composite laminate to maximize resistance against material failure and buckling across ten different loading scenarios.
• Modified File Structure & Modifications:
  • TASK_DETAILS.md & TASK_DETAILS_zh-CN.md: Registered the new task in the global task list.
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/README.md (and _zh-CN.md): Provided task overview, requirements, and execution commands.
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/Task.md (and _zh-CN.md): Detailed the engineering problem, physics (CLT, Ritz method), and scoring metrics.
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/scripts/init.py: Provided the entry point for the agent with a baseline solution (Haftka reference).
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/verification/: Implemented a robust evaluation suite including mechanics.py (physics engine), design_runtime.py (process isolation and timeout management), and evaluator.py.
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/references/: Included config.json for material properties/cases and design_notes.md for theoretical background.

2. AI Content Analysis

• Estimated AI Component: 15%
• Reasoning & Evidence: The core logic of the task—specifically the Classical Lamination Theory (CLT) implementation and the Ritz method for buckling—is highly domain-specific and derived from a cited scientific dataset (Zenodo DOI: 10.5281/zenodo.15864525). The design_runtime.py uses standard Python patterns for process isolation (e.g., subprocess.Popen, threading.Thread, queue.Queue), which might be AI-assisted but shows high-quality human architectural oversight. The documentation is professional and lacks the "fluff" or repetitive phrasing typical of pure AI generation.

3. Engineering & Economic Assessment

• Engineering Reality Check: This is a high-fidelity engineering problem. It moves beyond "toy" examples by incorporating real-world constraints (balanced/symmetric layups), material allowables, and complex physics (buckling vs. strain failure trade-offs). The use of process isolation in design_runtime.py to handle untrusted candidate code (timeouts, memory limits) reflects production-grade evaluation standards.
• Economic Value: High. Composite material optimization is a critical, high-cost domain in aerospace and automotive engineering. Automating the discovery of optimal stacking sequences can significantly reduce weight and material costs while maintaining safety margins.

4. Quality Assurance

• Verification & Testing:
  • frontier_eval Integration: Yes.
  • task_name: StructuralOptimization/CompositeLaminateStacking
  • Execution & Dependencies: The README.md clearly documents the installation of dependencies via verification/requirements.txt and provides specific commands for both direct and unified evaluation.
• Documentation Quality: Excellent. It includes bilingual documentation, clear API contracts, and a design_notes.md file that provides the mathematical and physical justification for the evaluator's logic.
• Organizational Structure: The structure is highly modular. Separating the physics (mechanics.py) from the execution environment (design_runtime.py) and the task configuration (config.json) makes the benchmark scalable and maintainable.

5. Security & Privacy Check

• Sensitive Files: Clean. No .env, API keys, or IDE-specific files were detected.
• Absolute Paths: None detected. The code uses Path(__file__) and relative paths for file operations, ensuring portability.

---

🇨🇳 中文分析

1. 摘要

• 核心目的: 此 PR 在 StructuralOptimization 类别下引入了一个名为 CompositeLaminateStacking 的新工程基准任务。该任务要求智能体优化 48 层复合材料层合板的铺层顺序，以在十种不同的载荷工况下最大化抗材料失效和抗屈曲能力。
• 修改的文件结构与变更摘要:
  • TASK_DETAILS.md & TASK_DETAILS_zh-CN.md: 在全局任务列表中注册了新任务。
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/README.md (及 _zh-CN.md): 提供了任务概览、环境要求和运行命令。
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/Task.md (及 _zh-CN.md): 详细说明了工程问题、物理模型（经典层合板理论 CLT、Ritz 法）和评分标准。
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/scripts/init.py: 为智能体提供了入口点及基准方案（Haftka 参考铺层）。
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/verification/: 实现了健壮的评测套件，包括 mechanics.py（物理引擎）、design_runtime.py（进程隔离与超时管理）和 evaluator.py。
  • benchmarks/StructuralOptimization/CompositeLaminateStacking/references/: 包含用于材料属性/工况的 config.json 和用于理论背景的 design_notes.md。

2. AI 成分分析

• 预估 AI 含量: 15%
• 判断依据与证据: 任务的核心逻辑——特别是经典层合板理论 (CLT) 的实现和用于屈曲分析的 Ritz 法——具有高度的领域专业性，并源自引用的科学数据集 (Zenodo DOI: 10.5281/zenodo.15864525)。design_runtime.py 使用了标准的 Python 进程隔离模式（如 subprocess.Popen, threading.Thread, queue.Queue），这可能经过 AI 辅助，但体现了高质量的人工架构监督。文档编写专业，没有纯 AI 生成常见的冗余或重复表述。

3. 工程与经济评估

• 工程现实检验: 这是一个高保真度的工程问题。它通过引入现实世界的约束（平衡/对称铺层）、材料许用值和复杂的物理特性（屈曲与应变失效之间的权衡），超越了“玩具”示例。在 design_runtime.py 中使用进程隔离来处理不可信的候选代码（超时、内存限制），反映了生产级的评测标准。
• 经济价值: 高。复合材料优化是航空航天和汽车工程中至关重要且成本高昂的领域。自动化发现最优铺层顺序可以在保持安全余量的同时，显著降低重量和材料成本。

4. 质量保证

• 验证与测试:
  • frontier_eval 集成: 是
  • task_name: StructuralOptimization/CompositeLaminateStacking
  • 运行与依赖: README.md 清晰地记录了通过 verification/requirements.txt 安装依赖的方法，并提供了直接评测和统一评测的具体命令。
• 文档质量: 优秀。包含双语文档、清晰的 API 契约，以及提供评测逻辑数学和物理依据的 design_notes.md 文件。
• 组织结构: 结构高度模块化。将物理计算 (mechanics.py)、执行环境 (design_runtime.py) 和任务配置 (config.json) 分离，使基准测试具备良好的可扩展性和可维护性。

5. 安全与隐私检查

• 敏感文件: 未发现异常。未检测到 .env、API 密钥或 IDE 特定文件。
• 绝对路径: 未检测到。代码使用 Path(__file__) 和相对路径进行文件操作，确保了可移植性。