feasibility-check

# 可行性检查 检查建模方案在给定数据和时间约束下是否可行: **$ARGUMENTS** ## Constants - REVIEWER_MODEL = `gpt-5.4` — Model used via Codex MCP. Must be an OpenAI model (e.g., `gpt-5.4`, `o3`, `gpt-4o`) ## Instructions Given a modeling approach description, systematically verify its feasibility: ### Phase A: 提取关键要素 1. Read the user's modeling approach description 2. Identify 3-5 core feasibility dimensions that need verification: - What mathematical method is proposed? - What data does it require? Does the available data match? - What is the computational complexity? - What is the implementation difficulty and time estimate? ### Phase B: 多源可行性检索 For EACH feasibility dimension, search using ALL available sources: 1. **Web Search** (via `WebSearch`): - Search for similar competition problems that used this method - Use specific technical terms from the approach - Try at least 3 different query formulations per dimension - Focus on: 数模竞赛优秀论文, 经典教材案例, 方法适用条件 2. **Known method references**: Check against: - Past CUMCM/MCM winning papers using similar methods - Classic textbook examples and applicable conditions 3. **Read details**: For each relevant reference, WebFetch its method section and results to assess real-world performance ### Phase C: 交叉模型验证 Call REVIEWER_MODEL via Codex MCP (`mcp__codex__codex`) with xhigh reasoning: ``` config: {"model_reasoning_effort": "xhigh"} ``` Prompt should include: - The proposed modeling approach - All references found in Phase B - Ask: "这个建模方案在给定数据和时间约束下可行吗？数据是否匹配？计算复杂度是否可控？实现时间是否合理？" ### Phase D: 可行性报告 Output a structured report: ```markdown ## 可行性检查报告 ### 建模方案 [1-2 sentence description] ### 可行性评估维度 1. [数据匹配度] — 可行性: HIGH/MEDIUM/LOW — 依据: [reference] 2. [计算复杂度] — 可行性: HIGH/MEDIUM/LOW — 依据: [reference] 3. [实现时间] — 可行性: HIGH/MEDIUM/LOW — 依据: [reference] 4. [方法适用性] — 可行性: HIGH/MEDIUM/LOW — 依据: [reference] ... ### 类似赛题案例 | 赛题 | 年份 | 竞赛 | 方法 | 效果 | |------|------|------|------|------| ### 综合可行性评估 - Score: X/10 - Recommendation: PROCEED / PROCEED WITH CAUTION / ABANDON - 关键优势: [what makes this approach feasible] - 风险点: [what could go wrong and how to mitigate] ### 改进建议 [How to adjust the approach to improve feasibility if needed] ``` ### Important Rules - Be BRUTALLY honest — an infeasible approach wastes precious competition time - "Standard method X applied directly" is feasible but may not score high — note the trade-off - Check both the method suitability AND the data compatibility - If the method is feasible but unlikely to score well, say so explicitly and suggest improvements - Always check whether the competition data actually supports the proposed method