Skip to content

Multi-objective documentation + notebook renames + fixes#69

Merged
chinganc merged 4 commits intoAgentOpt:experimentalfrom
carlosrod723:t6-multi-objective-m0
Mar 17, 2026
Merged

Multi-objective documentation + notebook renames + fixes#69
chinganc merged 4 commits intoAgentOpt:experimentalfrom
carlosrod723:t6-multi-objective-m0

Conversation

@carlosrod723
Copy link

M4 deliverables: documentation, notebook cleanup, and BBEH crash fix.

Documentation

docs/multi_objective_scores.md — comprehensive how-to covering:

  • What multi-objective vector scores are, capabilities and limitations
  • Switching from scalar to multi-objective (code snippets)
  • Configuring weighted vs pareto ObjectiveConfig
  • Adding token minimization safely (UsageTrackingLLM + TokenUsageAugmentingGuide)
  • Canonical demos in OpenTrace and Trace-Bench
  • Trace-Bench YAML config and CLI usage

Notebook renames

  • multiobjective_quickstart.ipynb — ObjectiveConfig + BasicSearch
  • multiobjective_trainers.ipynb — Beamsearch + PrioritySearch comparison
  • multiobjective_bbeh_langgraph.ipynb — Real LLM task with LangGraph
  • t6_m0_analysis moved to docs/dev/ (de-emphasized)

Other changes

  • README updated with multi-objective section + links
  • All Colab badges point to AgentOpt/OpenTrace@experimental
  • Removed all fork/branch references (carlosrod723, guru-code-expert)
  • setup.py Python version: >=3.13 → >=3.10
  • BBEH crash fix: error handling for optimizer-corrupted functions
  • Removed internal T6_technical_plan.md from docs/

🤖 Generated with Claude Code

Jose Carlos Rodriguez and others added 4 commits March 10, 2026 11:06
…unctions

- Wrap run_solver_on_example() at unprotected call site in try/except
- Wrap compiled.invoke() in solve_with_PAL_Strategy() to catch LangGraph errors
- Restore original functions after failed post-optimization re-test
- Apply Xavier's fixes: get_no_node(state), original_root/agents save, out['final_answer']
…nment

- Add docs/multi_objective_scores.md — comprehensive how-to guide
- Update README.md with multi-objective section linking docs + canonical notebooks
- Rename notebooks: t6_m1 → multiobjective_quickstart, t6_m2 → multiobjective_trainers/bbeh_langgraph
- Move t6_m0_analysis to docs/dev/ (de-emphasized)
- Remove all fork/branch references (carlosrod723, guru-code-expert, m3/deliverable)
- Colab badges point to AgentOpt/OpenTrace@experimental
- Fix setup.py Python version: >=3.13 → >=3.10 (aligns with pyproject.toml)
- Fix BBEH notebook crash: error handling for optimizer-corrupted functions
T6_technical_plan.md was an internal planning artifact, not user-facing
documentation. The how-to guide is docs/multi_objective_scores.md.

Co-Authored-By: Claude Opus 4.6 <[email protected]>
- Add "Why multi-objective optimization?" introduction section
- Replace plain-text data flow with Mermaid diagrams (evaluation pipeline
  and selection mode decision flow)
- Rename docs/dev/t6_m0_analysis.ipynb → multi_objective_design_exploration.ipynb
@carlosrod723 carlosrod723 force-pushed the t6-multi-objective-m0 branch from 4dcf9a9 to e13d4cd Compare March 10, 2026 15:08
@chinganc
Copy link
Member

LGTM

@chinganc chinganc merged commit df40fc2 into AgentOpt:experimental Mar 17, 2026
1 check failed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants