optimaru 0.1.0__tar.gz

optimaru-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

optimaru-0.1.0/MANIFEST.in

@@ -0,0 +1,13 @@
+# sdist for the `optimaru` Python distribution: the engine headers and, per use
+# case, the C++ sources its extension compiles plus the Python package.
+include LICENSE NOTICE pyproject.toml setup.py python/README.md
+recursive-include python/optimaru *.py
+recursive-include include *.hpp *.h
+recursive-include use-cases/*/cpp/include *.hpp *.h
+recursive-include use-cases/*/cpp/src *.cpp *.hpp *.h
+include use-cases/*/python/setup.py use-cases/*/python/bindings.cpp
+recursive-include use-cases/*/python *.py *.pyi py.typed
+prune use-cases/*/python/tests
+prune use-cases/*/python/build
+prune use-cases/*/python/.venv
+global-exclude *.so *.pyd __pycache__ *.pyc

optimaru-0.1.0/NOTICE

@@ -0,0 +1,10 @@
+optimaru
+Copyright 2026 The optimaru Authors
+
+This product's architecture, directory layout, class names and API are
+modelled on Timefold Solver (https://github.com/TimefoldAI/timefold-solver),
+which is licensed under the Apache License, Version 2.0.
+Copyright Timefold BV and contributors.
+
+optimaru is an independent C++ implementation and is not affiliated with or
+endorsed by Timefold BV. "Timefold" is a trademark of Timefold BV.

optimaru-0.1.0/PKG-INFO

@@ -0,0 +1,99 @@
+Metadata-Version: 2.4
+Name: optimaru
+Version: 0.1.0
+Summary: Metaheuristic optimization engine for routing, scheduling, timetabling and allocation — 21 ready-to-use models.
+Author: optimaru
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://optimaru.com
+Project-URL: Source, https://github.com/optimaru/optimaru
+Project-URL: Issues, https://github.com/optimaru/optimaru/issues
+Keywords: optimization,combinatorial-optimization,metaheuristics,scheduling,vehicle-routing,timetabling,operations-research,solver
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: MacOS
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Programming Language :: C++
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Classifier: Typing :: Typed
+Requires-Python: >=3.14
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: NOTICE
+Requires-Dist: pydantic>=2.5
+Dynamic: license-file
+
+# optimaru
+
+A metaheuristic optimization engine for real-world routing, scheduling,
+timetabling and allocation problems — with 21 ready-to-use models.
+
+The solver core is C++17 and ships compiled in the wheel; you use it from
+Python. Every business rule and preference is a constraint *plugin*: switch it
+on or off, change its weight, or make it hard or soft — without touching the
+solver.
+
+```bash
+pip install optimaru
+```
+
+Requires Python 3.14+. Wheels are published for Linux, macOS and Windows.
+
+## Example: vehicle routing
+
+```python
+from optimaru import vehicle_routing as vrp
+from optimaru.vehicle_routing.constraints import (
+    MinimizeTravelTime, get_constraints, maximize_visits_assigned, vehicle_capacity,
+)
+
+locations = [vrp.Location(37.5665, 126.9780)]          # 0: depot
+visits = []
+for i, (lat, lon) in enumerate([(37.5700, 126.9820), (37.5512, 126.9882), (37.5133, 127.1001)]):
+    locations.append(vrp.Location(lat, lon))
+    #            id       name           location demand  ready      due        service (s)
+    visits.append(vrp.Visit(f"C{i}", f"Customer {i}", i + 1, 2, 8 * 3600, 20 * 3600, 600))
+
+vehicles = [vrp.Vehicle("V1", 6, 0, departure_time=7 * 3600)]
+plan = vrp.VehicleRoutePlan("demo", locations, vehicles, visits)
+
+config = vrp.SolverConfig()
+config.spent_limit_seconds = 2
+solver = vrp.Solver(config)
+
+solver.solve(plan)                                      # default constraints
+print(plan.score, plan.routes())
+
+# Pick constraints and reweight them — the solver itself does not change.
+# Only the listed constraints are scored: leave out maximize_visits_assigned and
+# skipping every visit becomes free.
+solver.solve_with_constraints(plan, get_constraints([
+    vehicle_capacity,
+    maximize_visits_assigned,
+    MinimizeTravelTime(weight=10),
+]))
+print(plan.score, plan.routes())
+```
+
+## Use cases
+
+| Family | `optimaru.…` |
+| --- | --- |
+| Routing & logistics | `vehicle_routing`, `pickup_delivery`, `order_picking`, `facility_location` |
+| Scheduling | `flight_crew_scheduling`, `maintenance_scheduling`, `project_job_scheduling`, `semiconductor_scheduling`, `food_packaging`, `task_assigning`, `meeting_scheduling` |
+| Timetabling | `school_timetabling`, `conference_scheduling`, `sports_league_scheduling`, `tournament_scheduling` |
+| Allocation & packing | `bed_allocation`, `bin_packing`, `knapsack`, `assembly_line_balancing`, `quadratic_assignment`, `max_cut` |
+
+Each subpackage is loaded on first use, so `import optimaru` stays cheap.
+`list_constraints()` on a use case shows the constraint plugins it ships with.
+
+## Links
+
+- Website: <https://optimaru.com>
+- Source, docs and benchmarks against OR-Tools: <https://github.com/optimaru/optimaru>
+
+optimaru comes in a free, open-source edition (this package, Apache-2.0) and
+an Enterprise edition (coming soon).

optimaru-0.1.0/README.md

@@ -0,0 +1,242 @@
+# optimaru
+
+A dependency-free, header-only C++17 metaheuristic solver engine for
+real-world combinatorial optimization — scheduling, routing and assignment.
+
+The engine contains **no domain code** — problems consume it as a library.
+Worked use cases live in [`use-cases/`](use-cases) and public benchmark
+harnesses in [`benchmarks/`](benchmarks).
+
+optimaru comes in a free edition and a paid edition; this repository is the
+free, open-source edition.
+
+## Install (Python)
+
+```bash
+pip install optimaru
+```
+
+One package with every use case as a subpackage — `optimaru.vehicle_routing`,
+`optimaru.school_timetabling`, … — compiled for Linux, macOS and Windows
+(Python 3.14+). See [`python/README.md`](python/README.md) for an example.
+The distribution is built from this repository by [`setup.py`](setup.py), which
+reuses each use case's own `python/setup.py`.
+
+**Releasing:** bump `version` in [`pyproject.toml`](pyproject.toml) in a PR and
+merge it. On the push to `main`, [`wheels.yml`](.github/workflows/wheels.yml)
+builds the wheels and sdist and — only if there is no `v<version>` tag yet —
+uploads them to PyPI, then creates the `v<version>` tag and GitHub release.
+An unchanged version publishes nothing.
+
+## Repository layout
+
+```
+include/optimaru/     the engine (header-only, one header per class)
+test/                 engine tests
+use-cases/<slug>/     21 problems built on the engine: C++ domain + solver
+                      and a pybind11 package (or a small pure-Python model)
+benchmarks/<slug>/    the same problems against OR-Tools (CP-SAT, or the
+                      Routing solver for VRP) and, where available, Timefold;
+                      competing solvers live only here
+benchmarks/cpsat-common/  shared CP-SAT constraint-plugin contract
+```
+
+## Engine layout
+
+```
+include/optimaru/
+├── optimaru.hpp                       umbrella include for consumers
+├── api/
+│   ├── domain/solution_descriptor.hpp             planning variable description
+│   │                                              (VariableKind::BASIC | LIST | CHAINED)
+│   ├── domain/variable/variable_listener.hpp      shadow variables
+│   ├── score/buildin/hard_soft_score.hpp          HardSoftScore
+│   ├── score/calculator/incremental_score_calculator.hpp
+│   ├── score/calculator/chained_incremental_score_calculator.hpp
+│   ├── score/calculator/composite_incremental_score_calculator.hpp
+│   │                                              cost components, weighted sum
+│   ├── score/director/score_director.hpp          + EnvironmentMode::FULL_ASSERT
+│   ├── score/analysis/score_analysis.hpp          ScoreAnalysis / ConstraintAnalysis
+│   └── solver/{solver.hpp, solver_factory.hpp,
+│               solution_manager.hpp, solver_manager.hpp}
+├── config/                            SolverConfig as plain structs
+│   ├── solver/solver_config.hpp
+│   ├── solver/termination/termination_config.hpp
+│   ├── constructionheuristic/construction_heuristic_config.hpp
+│   ├── localsearch/local_search_config.hpp        + RuinRecreateConfig (ALNS)
+│   ├── localsearch/decider/acceptor/local_search_acceptor_config.hpp
+│   ├── heuristic/selector/move_selector_config.hpp  move mix weights (union selector)
+│   └── heuristic/selector/common/nearby/nearby_selection_config.hpp
+├── impl/
+│   ├── solver/default_solver.hpp                  phase orchestration, multi-start
+│   ├── solver/termination/termination.hpp         Termination hierarchy +
+│   │                                              SolverTermination (OR-composite)
+│   ├── constructionheuristic/default_construction_heuristic_phase.hpp
+│   ├── localsearch/default_local_search_phase.hpp step loop, sampled moves, LNS
+│   ├── localsearch/list_move.hpp                  the LIST moves every strategy
+│   │                                              previews / applies / undoes
+│   ├── localsearch/element_index.hpp              element -> (list, position)
+│   ├── localsearch/solution_snapshot.hpp          whole and list-wise restore
+│   ├── localsearch/strategy/                      Descent, EjectionChain,
+│   │                                              IteratedLocalSearch, ...
+│   ├── localsearch/decider/acceptor/              one class per subdirectory:
+│   │   ├── acceptor.hpp                           hillclimbing/ lateacceptance/
+│   │   │                                          tabu/ simulatedannealing/
+│   │   │                                          greatdeluge/ stepcountinghillclimbing/
+│   ├── heuristic/selector/common/nearby/nearby_neighborhoods.hpp
+│   └── heuristic/selector/move/generic/ruin_recreate_selector.hpp
+│                                                  LnsOutcome + RouletteWheelSelector
+├── preview/api/move/operators.hpp                 custom MoveOperator / RuinOperator
+│                                                  registration (experimental API)
+└── testing/preview_parity.hpp                     move-preview vs applied-score harness
+                                                   for a use case's calculator tests
+```
+
+`optimaru::` is one flat namespace; the directories organize the headers,
+the namespaces don't.
+
+## Features
+
+- **Planning variables**: BASIC (entity → value), LIST (ordered lists of
+  elements) and CHAINED (previous pointers/anchors, solved through a
+  synchronized list view so the full LIST move repertoire applies).
+- **Shadow variables** via `VariableListener<Solution>`
+  (`Solver::addVariableListener()`); FULL_ASSERT also catches stale shadows.
+- **Scoring**: `HardSoftScore` (int64 hard, double soft); constraints are
+  written as an `IncrementalScoreCalculator` (list hooks included) or composed
+  from named cost components with `CompositeIncrementalScoreCalculator`.
+  Move evaluation stays incremental (O(delta)).
+- **Score analysis**: `SolutionManager::analyze()` → `ScoreAnalysis` /
+  `ConstraintAnalysis`, a per-constraint breakdown over the named components.
+- **Verification**: `ScoreDirector` with `EnvironmentMode::FULL_ASSERT`;
+  score corruption throws `ScoreCorruptionError`.
+- **Construction heuristics**: First Fit, First Fit Decreasing, cheapest
+  insertion for list variables.
+- **Acceptors**: Hill Climbing, Late Acceptance, Entity Tabu, Simulated
+  Annealing, Great Deluge, Step Counting Hill Climbing.
+- **Terminations**: spent / unimproved-spent / step count /
+  unimproved-step-count / best score, OR-composed.
+- **Moves**: change/swap (BASIC); relocate / swap / 2-opt / Or-opt / 2-opt\*
+  tail swap (LIST), mixed by weight in `MoveSelectorConfig`.
+- **Nearby selection**: `NearbySelectionConfig::distanceMeter` biases moves
+  toward k-nearest neighbors.
+- **Ruin & recreate**: adaptive ALNS — random/related/worst ruin + greedy
+  (nearby-restricted) recreate + polish, roulette-wheel operator weights.
+- **Solving**: `SolverFactory` / `Solver` / `SolverConfig`; async background
+  solving with `SolverManager` / `SolverJob` / `SolverStatus` and
+  `terminateEarly()`; `solveMultiStart()` (N independent starts, best wins).
+- **Custom moves** (`preview/api/move`): `addMoveOperator` /
+  `addRuinOperator` / `setAcceptorFactory`.
+
+Multithreaded *move* evaluation is deliberately excluded: it only pays when a
+single score calculation is expensive (ms-scale); at this engine's ~200ns
+moves, per-step thread synchronization would cost more than it saves — use
+`solveMultiStart()` to occupy cores instead. Also out of scope:
+exhaustive/partitioned search and REST layers.
+
+## Using it
+
+Header-only: add `include/` to the include path, or from CMake:
+
+```cmake
+add_subdirectory(optimaru)   # or FetchContent
+target_link_libraries(my_app PRIVATE optimaru::optimaru)
+```
+
+```cpp
+#include "optimaru/optimaru.hpp"
+
+// 1. Describe the planning variable:
+template <>
+struct optimaru::SolutionDescriptor<MySolution> {
+    static constexpr optimaru::VariableKind kind = optimaru::VariableKind::BASIC;
+    static int entityCount(const MySolution&);
+    static int valueCount(const MySolution&);
+    static int getVariable(const MySolution&, int entity);   // -1 == unassigned
+    static void setVariable(MySolution&, int entity, int value);
+};
+// (LIST kind instead: listCount/elementCount/listSize/getElement/
+//  insertElement/removeElement/reverseSegment)
+
+// 2. Implement the constraints incrementally (or compose cost components
+//    with CompositeIncrementalScoreCalculator):
+class MyCalculator final : public optimaru::IncrementalScoreCalculator<MySolution> { ... };
+
+// 3. Configure and solve:
+optimaru::SolverConfig config;
+config.termination.spentLimit = std::chrono::seconds(30);
+config.localSearch.acceptor.acceptorType = optimaru::AcceptorType::LATE_ACCEPTANCE;
+config.localSearch.nearby.distanceMeter = [&](int a, int b) { return dist(a, b); };
+config.localSearch.ruinRecreate.enabled = true;   // ALNS kicks on stagnation
+
+auto solver = optimaru::SolverFactory<MySolution>::create(config).buildSolver();
+solver.addMoveOperator(std::make_unique<MyOrOptMove>());   // optional customs
+MyCalculator calculator;
+solver.solve(solution, calculator);
+```
+
+For a complete, production-shaped example see
+[`use-cases/vehicle-routing`](use-cases/vehicle-routing).
+
+## Performance features
+
+- **Nearby selection**: on spatial problems a distant relocate/swap is almost
+  never an improvement, so moves are biased toward k-nearest neighbors. The
+  same lists drive the LNS related ruin and the nearby-restricted repair.
+- **LNS / ALNS**: escapes the local optima that acceptor-only search cannot
+  leave — verified in the tests: hill climbing alone stays trapped on a
+  conflicted instance, hill climbing + LNS reaches the global optimum.
+- Measured on 100-city random TSP (5-seed sweep): nearby + LNS beats plain
+  local search on every seed even with HALF the outer step budget
+  (mean tour 2514 → 2409).
+
+## Use cases
+
+[`use-cases/README.md`](use-cases/README.md) lists all 21 use cases and
+explains how to extend one — adding a move, changing the construction
+heuristic, adding a hard or soft cost. Each use case's README has a short
+usage example and the same guide with its own file names.
+
+Scheduling: bed allocation, conference, flight crew, food packaging,
+maintenance, meeting, project-job (RCPSP), school timetabling, semiconductor
+fab, sports league, task assigning, tournament. Routing: vehicle routing
+(CVRPTW), pickup & delivery (PDPTW), order picking. Assignment and packing:
+facility location, bin packing, knapsack, quadratic assignment, max-cut,
+assembly-line balancing.
+
+```sh
+# C++
+cmake -S use-cases/<slug>/cpp -B build/<slug> && cmake --build build/<slug> -j
+ctest --test-dir build/<slug>
+
+# Python (uv)
+cd use-cases/<slug>/python && uv sync --all-groups --all-extras && uv run pytest
+```
+
+## Benchmarks
+
+[`benchmarks/README.md`](benchmarks/README.md) lists every comparison. In
+short: every use case is compared with OR-Tools CP-SAT on the same instances
+and rescored by the use case's own evaluator; vehicle routing is compared with
+the OR-Tools Routing solver and Timefold Solver on CVRPLIB Set X, and school
+timetabling three ways (engine / CP-SAT / Timefold). Pickup & delivery,
+project-job scheduling, bin packing, quadratic assignment and assembly-line
+balancing also run on public instance sets (Li & Lim, Patterson, Hexaly BPP,
+QAPLIB, SALBP).
+
+## Tests
+
+`test/engine_test.cpp` verifies every feature against brute force on two toy
+domains (BASIC task assignment, LIST TSP): all six acceptors, all four
+terminations, FFD ordering, multi-start, FULL_ASSERT catching a deliberately
+corrupted calculator, list moves under FULL_ASSERT, adaptive operator-weight
+convergence, LNS escaping a local optimum that traps hill climbing, and the
+nearby+LNS quality win on 100-city random TSP.
+
+```sh
+cmake -B build && cmake --build build && ctest --test-dir build
+```
+
+## License
+
+[Apache License 2.0](LICENSE). See [NOTICE](NOTICE) for attribution.