# SYN-LAYOUT-001 Rev A / Layout and material-flow method

**SYNTHETIC DEMONSTRATION — all facility dimensions, movement quantities, batches, service zones and costs are teaching assumptions. Results are calculations, not observed savings, employer work or a safety assessment.**

Factory SYN-OPS-01, operator SYN-021; product SYN-LC-1000; data version factory-1.0; layout version layout-1.0. Route: /engineering-lab/layout/.

## Problem, scope and responsibility

Investigate how a physical arrangement changes loaded material travel for one fictional cooling-skid product family. The visitor acts as a layout/improvement engineer: change an arrangement, inspect movement evidence, repair constraints and compare a feasible alternative. Seven areas and six planning transfers are modeled. Existing quality cases remain independent; a planning transfer assumes authorization before use rather than granting it.

The shared weekly demand is 1–24 whole skids. Editing demand creates a factory planning proposal. All compared arrangements use the same selected demand, BOM-derived quantities, batch capacities, rates and route rules. The three illustrative genealogy orders and original quality histories are not a weekly production schedule.

## Footprints, staging and utilities

Floor: 32 × 20 m. Position is the upper-left area origin, with x to the right and y downward. Footprints are unchanged from the factory baseline; sizes are not scaled by demand. Receiving R1 and dispatch D1 are fixed. Q1, M1, K1, A1 and T1 are movable, with kitting/staging contained in K1's footprint. M1 and T1 must remain entirely within their declared utility zones.

- SYN-SVC-M01 / Machining electrical/service zone: x=0, y=8, width=10, depth=12 m.
- SYN-SVC-T01 / Test utility zone: x=24, y=8, width=8, depth=12 m.

No footprints may overlap or extend outside the floor. Each pair must have at least 2 m separation on at least one axis. This conservative orthogonal geometry rule reserves the assumed clearance; it does not certify real circulation, fire access, maintenance access or a vehicle turning envelope.

## From-to transfers and batch assumptions

| Movement | From / to | Units per skid | Capacity per loaded trip | Basis |
| --- | --- | --- | --- | --- |
| SYN-MOVE-01 | R1 / Q1 | 1 harness sets | 4 | One set per skid, four sets per loaded cart trip |
| SYN-MOVE-02 | Q1 / K1 | 1 harness sets | 4 | Same four-set cart; assumes disposition is authorized before movement |
| SYN-MOVE-03 | M1 / K1 | 1 valves | 4 | One valve per skid, four valves per loaded bin trip |
| SYN-MOVE-04 | K1 / A1 | 1 kits | 1 | One complete kit per skid, one kit per loaded trip |
| SYN-MOVE-05 | A1 / T1 | 1 skids | 1 | One skid per loaded transfer; no dimensional vehicle model |
| SYN-MOVE-06 | T1 / D1 | 1 skids | 1 | One skid per loaded transfer; assumes applicable reviews are complete |

For each movement, loaded trips = ceiling(demand × units per skid / batch capacity). Partial batches still need a trip. The matrix in the workspace shows these directional counts. No extra movement is inferred for unspecified purchased components; kit transfers stand for the complete order kit. Empty returns, inbound vehicles, within-area handling and rework are excluded.

## Routing method

Transfer points lie 1 m beyond each declared top/bottom face, at the rounded integer horizontal centre. Points represent aisle handoff locations, not machine centroids. A pair can share a handoff point, producing zero aisle distance; dispatch effort remains in the per-trip cost, while within-area maneuvering is excluded.

Grid points are spaced 1 m apart; valid centres remain at least 1 m from the floor boundary. Each rectangular footprint is expanded 1 m on all sides. Grid centres inside the expanded footprint are blocked; boundary centres are permitted. Horizontal/vertical steps of length 1 m connect the remaining centres. No diagonal step cuts a corner.

A four-neighbour breadth-first search finds a shortest route on this unweighted grid, using deterministic east/north/west/south tie-breaking. Route distance is the number of edges. This is a grid shortest path, not a continuous-distance or vehicle-motion optimum. The drawn spaghetti paths use these actual route points, with line width increasing with loaded-trip count.

If a port or path is unavailable, the arrangement fails feasibility. Geometry, fixed-area, utility-zone and route violations are reported explicitly. An infeasible arrangement has no aggregate travel/cost score or claimed improvement; individual reachable route records remain inspectable for diagnosis.

## Objective and cost

Loaded travel = sum(route metres × loaded trips). Trip counts supply the objective weights. Handling cost = loaded metres × $0.08 + loaded trips × $0.40. These composite rates are invented teaching assumptions, not measured wages, cycle times or quotations. Fixed trip cost is identical across arrangements at the same demand; minimizing travel therefore minimizes the geometry-dependent cost in this model.

- **Loaded travel / metres / hypothetical week**: Sum of shortest grid route distance × loaded trips for the six declared movements. Excludes empty returns and within-area movement.
- **Loaded trips / trips / hypothetical week**: For each movement: ceiling(requested skids × units per skid / batch capacity). Partial batches still require a trip.
- **Assumed handling cost / USD / hypothetical week**: $0.08 per loaded metre plus $0.40 per loaded trip; fictional composite cost assumptions, without a wage or time forecast.
- **Change from baseline / metres and USD / hypothetical week**: Baseline minus candidate, with identical demand, batches, costs and route rules. Positive values are hypothetical reductions. Infeasible arrangements are not compared.

## Bounded heuristic and decisions

Search checks the unchanged baseline, a feasible visitor proposal, two coordinated central-aisle shifts, a coordinated utility-zone packing start and a same-footprint Q1/K1 swap. It then performs at most three passes of four 1 m neighbour moves for each movable area. Only feasible candidates that strictly reduce weighted loaded metres replace the incumbent. The baseline is always included; the default result cannot be worse than it. The declared maximum is 66 evaluated arrangements, with deterministic ordering and no randomness.

This is a feasible heuristic improvement, not a global optimum. Search records include demand, evaluated count and strategy. Demand changes re-score all arrangements using current demand; an older suggestion is marked for regeneration. Selecting a view preserves the other arrangements. Editing any view forks the displayed geometry into a visitor proposal. Use suggestion as proposal adopts its geometry without changing any quality review.

## Reproducible example / 8 skids

Baseline route lengths: SYN-MOVE-01: 8 m × 2 trips; SYN-MOVE-02: 8 m × 2 trips; SYN-MOVE-03: 17 m × 2 trips; SYN-MOVE-04: 5 m × 8 trips; SYN-MOVE-05: 11 m × 8 trips; SYN-MOVE-06: 2 m × 8 trips. Total 210 loaded m and 30 trips; assumed cost $28.80 per hypothetical week. The default search evaluates 26 arrangements and returns 170 loaded m / $25.60, a calculated reduction of 40 m. These values come from the implementation and are checked against independent arithmetic and route cases.

## Implementation and export

data/factory/layout.ts declares geometry/transfer assumptions. lib/factory/layout.ts provides constraint checks, grid routing, scoring and the bounded heuristic. lib/factory/layout-session.ts handles versioned layout state, validation and comparison exports. LayoutWorkbench provides pointer, keyboard and coordinate controls. The factory provider persists browser-tab state; schema 1 sessions migrate to schema 2 with their planning and quality states preserved and a fresh baseline layout.

Reset layout restores only layout geometry/views and preserves demand plus SPC/NCR/FAT reviews. Reset factory restores everything. Export layout comparison includes assumptions, all arrangements, route points, trip counts, violations, costs and differences. The larger factory session export includes the same layout comparison alongside independent quality states. No backend, confidential data, actual release or AI model is involved.

Regenerate with pnpm docs:layout. Verify with pnpm check:layout. Inspect layout-manifest.json for machine-readable assumptions.

## Results and lessons

The visitor's chosen feasible arrangement demonstrates how batch weighting, transfer-point locations and fixed utilities affect movement. A geometric improvement does not establish a capacity, ergonomic or quality improvement. The model omits variability, congestion, replenishment policies, turns, material handling speeds and returns; later phases add distinct assessments rather than converting this arithmetic into measured operating results.