Add Dembo geometric programming test problems from 1976 paper

[arnavk23]

Towards #284

[Copilot]

Pull request overview

Adds a new family of Dembo (1976) geometric-programming-inspired benchmark problems to the library, aiming to make them available in both the PureJuMP and ADNLPProblems backends with corresponding entries in the global metadata table.

Changes:

• Added new PureJuMP model constructors for dembo_gp1a–dembo_gp8a (subset).
• Added new ADNLPProblems constructors for the same names.
• Added src/Meta/* entries intended to register these problems in OptimizationProblems.meta.

Reviewed changes

Copilot reviewed 30 out of 30 changed files in this pull request and generated 30 comments.

Show a summary per file

File	Description
src/PureJuMP/dembo_gp1a.jl	Adds JuMP model for Dembo GP1A (docstring includes a solution section).
src/PureJuMP/dembo_gp1b.jl	Adds JuMP model for Dembo GP1B variant.
src/PureJuMP/dembo_gp2.jl	Adds JuMP model for Dembo GP2.
src/PureJuMP/dembo_gp3.jl	Adds JuMP model for Dembo GP3.
src/PureJuMP/dembo_gp4a.jl	Adds JuMP model for Dembo GP4A.
src/PureJuMP/dembo_gp4c.jl	Adds a JuMP model file intended for Dembo GP4C (currently misnamed/mis-exported).
src/PureJuMP/dembo_gp5.jl	Adds JuMP model for Dembo GP5.
src/PureJuMP/dembo_gp6.jl	Adds JuMP model for Dembo GP6.
src/PureJuMP/dembo_gp7.jl	Adds JuMP model for Dembo GP7.
src/PureJuMP/dembo_gp8a.jl	Adds JuMP model for Dembo GP8A.
src/ADNLPProblems/dembo_gp1a.jl	Adds ADNLPModel for dembo_gp1a (currently inconsistent/infeasible).
src/ADNLPProblems/dembo_gp1b.jl	Adds ADNLPModel for dembo_gp1b (currently inconsistent with JuMP variant).
src/ADNLPProblems/dembo_gp2.jl	Adds ADNLPModel for dembo_gp2 (currently a different problem than JuMP).
src/ADNLPProblems/dembo_gp3.jl	Adds ADNLPModel for dembo_gp3 (currently a different problem than JuMP).
src/ADNLPProblems/dembo_gp4a.jl	Adds ADNLPModel for dembo_gp4a (objective sign differs vs JuMP).
src/ADNLPProblems/dembo_gp4c.jl	Adds an ADNLP model file intended for Dembo GP4C (currently misnamed/mis-exported).
src/ADNLPProblems/dembo_gp5.jl	Adds ADNLPModel for dembo_gp5 (currently a different problem than JuMP).
src/ADNLPProblems/dembo_gp6.jl	Adds ADNLPModel for dembo_gp6 (missing positivity bounds despite fractional powers).
src/ADNLPProblems/dembo_gp7.jl	Adds ADNLPModel for dembo_gp7 (currently a different problem than JuMP).
src/ADNLPProblems/dembo_gp8a.jl	Adds ADNLPModel for dembo_gp8a (currently a different problem than JuMP).
src/Meta/dembo_gp1a.jl	Registers metadata for dembo_gp1a (currently invalid :origin type and inconsistent classification).
src/Meta/dembo_gp1b.jl	Registers metadata for dembo_gp1b (currently invalid :origin type).
src/Meta/dembo_gp2.jl	Registers metadata for dembo_gp2 (currently invalid :origin type).
src/Meta/dembo_gp3.jl	Registers metadata for dembo_gp3 (currently invalid :origin type).
src/Meta/dembo_gp4a.jl	Registers metadata for dembo_gp4a (currently invalid :origin type and misclassified lin/nonlin split).
src/Meta/dembo_gp4c.jl	Registers metadata for a GP4* problem (currently name/variable mismatch vs filename; will break module init).
src/Meta/dembo_gp5.jl	Registers metadata for dembo_gp5 (currently invalid :origin type and misclassified constraint type).
src/Meta/dembo_gp6.jl	Registers metadata for dembo_gp6 (currently invalid :origin type).
src/Meta/dembo_gp7.jl	Registers metadata for dembo_gp7 (currently invalid :origin type).
src/Meta/dembo_gp8a.jl	Registers metadata for dembo_gp8a (currently invalid :origin type and misclassified as nonlinear constraints).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 31 out of 31 changed files in this pull request and generated 36 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[arnavk23]

@tmigot Whenever you are free, please take a look at this pr. For now I have not implemented the docstring for these problems, whatever we all agree in your pr, I will follow and add a commit for that later.

[arnavk23]

AI-Assisted Problem Generation Workflow

Model & Tools Used

• Model used in this run: GPT-5.3-Codex
• Code tools: grep_search, file_search, read_ofile, replace_string_in_file, multi_replace_string_in_file, get_errors, fetch_webpage

Prompts (2 iterations)

Prompt 1

"Implement the problems from the pdf. I want everything done correctly, 
if not able to do better to remove than to add incorrect implementations. 
Check all the problems you added with the pdf then try to add the missing ones. 
If incorrect try to fix. If not able to fix, remove."

Context provided:

• PDF file: BF01580667.pdf (Dembo 1976 paper, ~22 pages)
• Three backends: ADNLPProblems, PureJuMP, Meta

AI Response Process:

1. Used fetch_webpage to read PDF content (limited effectiveness on scanned PDFs, so later used pdftotext)
2. Read existing implementations with read_file to understand codebase patterns
3. Identified: 8 problems described in paper, only some had readable tables

Prompt 2

"leave the ones you think can't be done properly as I can't 
provide more than this"

AI Response:

• Decided to implement only GP2-GP7 (6 problems)
• Removed GP1A, GP1B, GP8A (OCR-corrupted exponents in tables)
• Kept strict fidelity constraint rather than guessing

Iterative Refinement Loop

Iteration 1: Initial Draft

AI Action:
1. Search for problem descriptions in PDF using grep
2. Extract coefficient values manually from tables
3. Generate ADNLPProblems template code
4. Generate PureJuMP wrapper code
5. Generate Meta metadata dicts

Iteration 2: Fix Syntax

AI Action:
1. Use get_errors() to identify syntax problems
2. Use multi_replace_string_in_file to remove concatenated legacy code
3. Re-validate with get_errors()

Iteration 3: Test & Cache Issues

AI Action:
1. Run Pkg.test() → fails with "list_problems ≠ meta[!, :name]"
2. Ran comprehensive metadata check

How to Reproduce with API

Using Claude API with Problem Specification

Example request structure:

import anthropic
import json

client = anthropic.Anthropic()

problem_spec = {
    "name": "dembo_gp2",
    "source": "Dembo 1976, Table 2.1",
    "nvar": 5,
    "ncon": 6,
    "objective": "c[1]*x[2] + c[2]*x[1]*x[5] + ...",
    "coefficients": [1.4, 0.8, ...],
    "bounds": {"lower": [78, 33, 27, 27, 27], "upper": [102, 45, 45, 45, 45]},
    "constraints": [
        {"type": "rational", "formula": "c[11]/(x[2]*x[5]) + ... ≤ 1"}
    ]
}

prompt = f"""
Generate three implementations of this optimization problem.

Problem specification (JSON):
{json.dumps(problem_spec, indent=2)}

Templates to use:
1. ADNLPProblems template:
[ADNLPProblems_template_code]

2. PureJuMP template:
[PureJuMP_template_code]

3. Meta template:
[Meta_template_code]

Output format:
1. /src/ADNLPProblems/{problem_spec['name']}.jl
2. /src/PureJuMP/{problem_spec['name']}.jl
3. /src/Meta/{problem_spec['name']}.jl

Validate each against the problem specification JSON.
"""

response = client.messages.create(
    model="claude-3-5-haiku-20241022",
    max_tokens=4096,
    messages=[{"role": "user", "content": prompt}]
)

print(response.content[0].text)

Iterative Validation Loop (API)

def validate_and_refine(code_files, spec):
    """Iteratively validate and fix code."""
    
    for iteration in range(max_iterations):
        # Check syntax
        errors = check_syntax(code_files)
        
        if not errors:
            # Verify against spec
            consistency = check_spec_compliance(code_files, spec)
            if consistency == "pass":
                return code_files, "success"
        
        # Request fix from Claude
        prompt = f"""
        Found issues in generated code:
        {errors}
        
        Problem spec:
        {json.dumps(spec)}
        
        Files:
        {code_files}
        
        Fix these issues while maintaining the problem specification.
        """
        
        response = client.messages.create(...)
        code_files = extract_code_from_response(response)
    
    return None, "max_iterations_reached"

Tools Used & Their Purpose

Tool	Purpose	Example
grep_search	Find existing implementations to understand patterns	Search for "export dembo"
read_file	Understand codebase structure	Read ADNLPProblems/aircrfta.jl as template
replace_string_in_file	Fix targeted syntax issues	Remove concatenated legacy code
multi_replace_string_in_file	Batch fixes across multiple files	Fix 7 files simultaneously
get_errors()	Validate code after changes	Confirm no remaining syntax errors
fetch_webpage	Attempt PDF content extraction	Initial exploratory read of PDF