Fix #506: Add OpenShopScheduling model

[isPANN]

No description provided.

[isPANN]

Implementation Summary

Changes

• src/models/misc/open_shop_scheduling.rs — New OpenShopScheduling model: n×m direct-permutation config, greedy simulation evaluator, Value = Min<usize>, declare_variants! with factorial(num_jobs)^num_machines complexity, canonical example with 4×3 instance
• src/models/misc/mod.rs — Added module declaration and pub re-export for OpenShopScheduling
• src/rules/openshopscheduling_ilp.rs — Direct ILP reduction using disjunctive formulation: binary x_{j,k,i} (job ordering per machine), binary y_{j,i,i'} (task ordering per job), integer s_{j,i} (start times), integer C (makespan objective)
• src/rules/mod.rs — Added feature-gated module declaration and canonical example registration
• src/unit_tests/models/misc/open_shop_scheduling.rs — 19 unit tests (creation, dims, evaluate, decode_orders, compute_makespan, serialization, brute-force)
• src/unit_tests/rules/openshopscheduling_ilp.rs — 6 ILP tests (structure, closed-loop, extract_solution, edge cases)
• docs/paper/reductions.typ — Added problem-def("OpenShopScheduling") with Gantt chart and reduction-rule("OpenShopScheduling", "ILP") with full formulation
• docs/paper/references.bib — Added Gonzalez & Sahni 1976 JACM entry

Deviations from Plan

• Canonical example optimal value: The issue body claimed optimal makespan = 11 for the 4×3 instance. Exhaustive brute-force reveals the true optimal is 8 (224/13824 orderings achieve it; the issue's schedule with makespan 11 is valid but suboptimal). The canonical example stores the correct optimal config and value.
• Upper bounds on ILP start times: Added s_{j,i} <= total_p bounds to prevent unbounded integer search; without these, the ILP solver timed out on even 2×2 instances.

Open Questions

• None.

[isPANN]

Closing - will be included in batch-models branch instead

[codecov]

Codecov Report

❌ Patch coverage is 99.39811% with 14 lines in your changes missing coverage. Please review.
✅ Project coverage is 97.94%. Comparing base (a70b132) to head (54a63d1).

Files with missing lines	Patch %	Lines
.../rules/sequencingwithdeadlinesandsetuptimes_ilp.rs	96.03%	4 Missing ⚠️
src/types.rs	0.00%	3 Missing ⚠️
src/models/misc/preemptive_scheduling.rs	98.44%	2 Missing ⚠️
...ls/misc/minimum_internal_macro_data_compression.rs	99.00%	1 Missing ⚠️
src/models/misc/open_shop_scheduling.rs	99.00%	1 Missing ⚠️
...c/rules/minimuminternalmacrodatacompression_ilp.rs	99.38%	1 Missing ⚠️
src/unit_tests/rules/openshopscheduling_ilp.rs	98.64%	1 Missing ⚠️
.../rules/sequencingwithdeadlinesandsetuptimes_ilp.rs	98.71%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #959      +/-   ##
==========================================
+ Coverage   97.89%   97.94%   +0.04%     
==========================================
  Files         651      671      +20     
  Lines       71463    73707    +2244     
==========================================
+ Hits        69959    72189    +2230     
- Misses       1504     1518      +14     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.