Purpose

Generate a candidate method pool for each classified subquestion in a mathematical modeling contest.

This skill converts validated problem parse, problem classification, and related-paper analysis into a structured method candidate pool. For each subquestion, it proposes 2-4 meaningfully different candidate modeling schemes, compares them on feasibility, interpretability, data fit, and implementation risk, and recommends a first-round execution priority. It specifies what each method expects as input, what it should output, and how its results should be evaluated.

This skill focuses on generating the candidate pool — the actual method selection happens AFTER multi-round experiments, when the modeler reviews experiment reports and locks in the final method.

This skill does not generate code, clean data, run experiments, create figures, or write paper sections.

When to use

Use this skill:

• After problem-parser, problem-classifier, and related-paper-analyzer have produced validated artifacts.
• Before data cleaning, code generation, robustness checks, figure planning, or paper writing.
• When the modeler needs a structured set of candidate methods to consider for each subquestion.
• When the user says: "what methods should we try for Q1?", "generate candidate methods", "build the method pool for Q2".
• When the team needs to see multiple modeling approaches before committing to one.