Quickstart

This example shows the shortest meaningful path through design-research-problems.

Requires Python 3.12+.

Note

For an editor-first setup, interpreter selection, and debugging flow, see Run An Example In VS Code.

1. Install

python -m pip install design-research-problems

On Windows, if python resolves to an older interpreter, use py -3.12 -m pip install design-research-problems and py -3.12 -m venv .venv.

Or install from source:

git clone https://github.com/cmudrc/design-research-problems.git
cd design-research-problems
python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install -e .

2. Minimal Runnable Example

import design_research_problems as derp

print(f"Catalog size: {len(derp.list_problems())}")
summaries = derp.search_problem_summaries(text="pill", kind="optimization")
print(summaries[0].to_dict())

problem = derp.get_problem("ideation_peanut_shelling_fu_cagan_kotovsky_2010")
print(problem.render_brief(include_citation=False))

optimization = derp.get_problem("pill_capsule_min_area")
if isinstance(optimization, derp.OptimizationProblem):
    print(optimization.solver_hints())

3. What Happened

You loaded one packaged study task from the catalog and rendered its design brief. This is the core pattern for building comparable task inputs before adding evaluators, solvers, or orchestration.

search_problem_summaries returns compact metadata for selection and routing without putting every full problem brief into an agent context window. Optimization problems also expose solver_hints() so callers can see bounds, variable domain, constraint counts, and a solver-family hint directly.

4. Where To Go Next

• Concepts

• Typical Workflow

• Problems

• Examples

• API

Ecosystem Note

In a typical study, design-research-agents provides executable participants, design-research-problems supplies the task, design-research-experiments defines and orchestrates the study, and design-research-analysis interprets the resulting records.