Algorithms

Algorithms

The orchestration layer that strings together the Haldane / Schreiner equations, Bühlmann M-values, and the Baker gradient-factor ramp into a working dive simulator. The Decompression-Model chapters cover the underlying equations; this section is about how those equations are called to produce a dive plan.

Pipeline

1. The user provides a dive setup: bottom gas + optional deco gases, $GF_{low}$ / $GF_{high}$, and either a full waypoint array or just {maxDepth, bottomTime}.
2. generateDecoProfile(maxDepth, bottomTime, gases, gfLow, gfHigh, …) (js/diveSetup.js:326) is the top-level entry. It calls calculateNDL() first; if $t_{bottom} \le NDL$ it returns a simple no-stop profile. Otherwise it simulates descent + bottom time to get tissue state, then hands off to generateDecoSchedule().
3. generateDecoSchedule() (js/decoModel.js) computes gas-switch MODs, calls findFirstStopAtGFLow() to determine pAnchor and the first stop in one pass, then runs the deco loop to generate the stop list.
4. Output is a waypoint array (embedded stops + gas switches) plus metadata (total deco time, controlling compartment, pAnchor, anchorDepth).
5. Chart rendering replays the waypoints at 10-second resolution via calculateTissueLoading() (js/decoModel.js:1178) and overlays per-timepoint ceilings via calculateCeilingTimeSeriesDetailed() (js/decoModel.js:617).

flowchart TD
  A[generateDecoProfile<br/>diveSetup.js:326] --> B{bottomTime<br/>vs NDL}
  B -- "t ≤ NDL" --> C[generateSimpleProfile]
  B -- "t > NDL" --> D[simulate descent + bottom]
  D --> E[generateDecoSchedule<br/>decoModel.js:899]
  E --> F[findFirstStopAtGFLow<br/>strict GF_low first-stop search]
  F --> H[deco stop loop]
  H --> I[waypoints + stops]
  I --> J[calculateTissueLoading<br/>decoModel.js:1178]
  I --> K[calculateCeilingTimeSeriesDetailed<br/>decoModel.js:617]

Loading

The pipeline is single-pass per dive setup: tissue state flows forward from surface, pAnchor is computed once at ascent start and passed downstream to every function that needs a GF-ramp reference.

Chapter TOC

1. Algo-01-Ascent-Simulation — replay a waypoint array over all 16 compartments; segment dispatch between Haldane (level) and Schreiner (slope).
2. Algo-02-NDL-Calculation — binary search for no-decompression limit at a given depth and gas.
3. Algo-03-First-Stop-Ramped-GF — the two-step pAnchor → first-stop discovery. The subtlest chapter.
4. Algo-04-Deco-Stop-Loop — from first stop to surface, generate the stop list; the DECO_STOP_MAX_MINUTES safety cap.
5. Algo-05-Multi-Gas-Switching — MOD calculation, sequential switching, gas-switch waypoint insertion.
6. Algo-06-Ceiling-Time-Series — per-timepoint ceiling overlay for chart rendering.

Cross-cutting: Architecture for the module/function index; Decompression-Model for the equations themselves; References for the prior-art decotengu docs whose structure these chapters mirror.