# Combination pipelines

> The named search experiments that chase score. Each is a pipeline rather than a single algorithm: it builds a board with one engine, then lifts or finishes it with another. Each records its idea, its best board, and the questions it left open.

- Canonical page (with interactive figures/demos): https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/
- Updated: 2026-07-16

---
Every page here is one named experiment with an idea and a board, and nearly
every one is a *pipeline*: it builds a board with one engine, then lifts or
finishes it with another. That is what sets this group apart from the two studies
that take a single paradigm apart, the
[DFS study](/research/lab/experiments/raphael-anjou/dfs-study) and the
[repair study](/research/lab/experiments/raphael-anjou/repair-study). These chain
engines together, and the interest
is as much in the *combination*, the division of labour between construction,
repair and an exact endgame, as in any one stage.

They share the [engines](/research/lab/experiments/raphael-anjou/engines); what
differs is how each one composes and steers them: what it seeds, what it forbids,
what it tears up and rebuilds. Each page's method section walks its stages in
order. Several of the from-scratch builders lean on the beam producer, and two on
the ALNS repair loop, engines that do not yet have their own write-ups here (see
the [engines page](/research/lab/experiments/raphael-anjou/engines)); where a
pipeline does, its page says so plainly rather than pointing at a page that is not
up yet.

They are ordered by what they taught, not by score. A pipeline that ended lower
but explained why is worth more here than one that scraped a point and could not
say how.

## Pages in this section

- [GAUNTLET](https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/gauntlet) — Run the same beam search across nine different scan orders, so it lands in different regions instead of always converging to the same one. The zigzag order found a brand-new 458 board.
- [CLOISTER](https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/cloister) — Fix a perfect border, then search the interior with the border's edges treated as hard constraints from the very first cell.
- [MIDDEN](https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/midden) — Decide in advance not when a board may break, but where: confine every mismatch to a chosen shape of cells, and search for the best shape.
- [LADDER](https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/ladder) — Throw hundreds of cheap short searches at the board, keep only the deepest starts, and promote the survivors through longer and longer rounds.
- [MOSAIC](https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/mosaic) — Tile the board into small blocks, solve each one to proven optimality, and glue them together, paying for the seams instead of forbidding them. From scratch, with no record to copy, it reaches 448.
- [STAGED](https://eternity2.dev/research/lab/experiments/raphael-anjou/pipelines/staged) — Build the whole board from scratch with no pre-set frame, in stages, letting the border emerge last from whatever pieces are left.