# Substudy NN: [Descriptive Name]

**Date**: YYYY-MM-DD
**Status**: [planned | running | completed | failed]
**Trials**: N

## Purpose

[Why was this substudy created? What hypothesis is being tested?]

Example:
- Validate parameter update mechanism
- Explore promising design region identified in previous run
- Test relaxed constraint settings
- Sensitivity analysis for specific parameter

## Configuration Changes

**From Previous Substudy** (or baseline):

[List what changed compared to previous substudy or baseline config]

Examples:
- Increased beam thickness bounds from 40mm to 60mm
- Changed displacement constraint from 10mm to 12.5mm
- Adjusted objective weights: displacement 50% (was 33%), stress 25% (was 33%), mass 25% (was 34%)
- Narrowed hole_count range to 9-13 (was 5-15)
- Increased trial count from 10 to 50

**Design Variables**:
- `var1`: min-max units [if changed, note it]
- `var2`: min-max units

**Objectives**:
- objective1: weight [if changed, note it]
- objective2: weight

**Sampler**: [TPE | Random | Grid | etc.]

## Expected Outcome

[What do you hope to learn or achieve?]

Examples:
- Find feasible designs meeting relaxed constraint
- Identify which parameter has most impact on displacement
- Validate that hole_count updates correctly
- Achieve 10% improvement over best result from substudy 03

## Actual Results

[Fill in after completion]

**Status**: [✅ SUCCESS | ⚠️ PARTIAL | ❌ FAILED]

**Best Trial**: #XX
- Objective: XXX
- Key metrics: [e.g., displacement: X.X mm, mass: XXXX kg]
- Feasible: [Yes/No]

**Key Findings**:
- Finding 1
- Finding 2
- Finding 3

**Statistics**:
- Total trials: XX
- Feasible designs: N / N_total (XX%)
- Best objective: XXX
- Improvement over previous: XX%

## Validation

[If this was a validation run, document what was validated]

Examples:
- ✅ All design variables updated correctly
- ✅ FEA results reflect parameter changes
- ✅ Mesh regenerated properly
- ✅ Expression import via .exp method works

## Lessons Learned

[What did we learn that wasn't expected?]

Examples:
- Optimizer consistently pushes beam thickness to maximum bounds
- hole_count has minimal impact on displacement (< 5% variation)
- Constraint appears infeasible with current design variable ranges
- TPE sampler converges after ~30 trials

## Visualization

[Link to generated plots if available]

- [convergence.pdf](plots/convergence.pdf)
- [design_space_evolution.pdf](plots/design_space_evolution.pdf)
- [parallel_coordinates.pdf](plots/parallel_coordinates.pdf)

## Next Steps

[What should be done next based on these results?]

Examples:
- → Substudy 05: Run refined search with narrowed parameter bounds
- → Increase beam thickness limit to 80mm to meet constraint
- → Perform sensitivity analysis to identify most important parameters
- → Study complete - recommend design #43 for detailed validation

---

**Files**:
- Configuration: `config.json`
- Results: `history.json`, `best_trial.json`
- Detailed Analysis: `OPTIMIZATION_RESULTS.md` (if significant findings)