Atlverse Simulation

Atlverse Simulation is a discrete-event simulation engine written in Python (package simcore) with a web editor on top. The engine implements the SIMAN modelling language — entities flowing through blocks, resources and queues, schedules, transfer (stations, routes, transporters, conveyors, guided vehicles), combined discrete/continuous models and batch-means statistics — and the editor offers the familiar Basic Process, Advanced Process and Advanced Transfer modules. The whole thing also runs inside the browser (Pyodide), so a class can use it from a link, each student's laptop doing its own simulation.

Results are statistically comparable to Arena, not digit-for-digit identical (event ordering on ties differs). COVERAGE.md lists every SIMAN block and element and its status.

Features

Runs anywhere — a local server (python -m simcore.web) or a static site (python -m simcore.web.build) for GitHub Pages, Cloudflare Pages, Netlify or a campus web server; works offline after the first visit.

Quick start

In the browser: open https://simulation.atlverse.xyz, or host your own copy (see In the browser).

On your computer (Python 3.10+, no other packages needed; PyYAML only for YAML files):

git clone https://github.com/andrewtliem/atlverse-simulation.git
cd atlverse-simulation
python -m simcore.web            # opens http://127.0.0.1:8765, models saved in ./models

Example models from Altiok & Melamed, Simulation Modeling and Analysis with Arena, chapter 5 (5-2 workstation, 5-6 two processes, 5-7 emergency room) are under File → Open….

Three ways to build a model

1. Web editor — as above. Shortcuts: Del, Ctrl+Z/Y, Ctrl+C/V/D, Ctrl+A, Ctrl+S, F5 (Go), F10 (Step), Esc (Pause), Shift+F5 (End), F1 (Help).

2. YAML — the engine's own model format (simcore/loader.py documents every key):

python -m simcore examples/bank.yaml --reps 10

3. Python API:

from simcore import *

m = Model("M/M/1", time_units="minutes")
m.resource("Server")
m.add(Create(interval="EXPO(1)"), Queue("Server.Queue"), Seize(resources=[("Server", 1)]),
      Delay(duration="EXPO(0.8)"), Release(resources=[("Server", 1)]), Dispose())
m.replicate(replications=5, length=10_000)
res = m.run()
print(res.report())
print(res.across("queues.Server.Queue.waiting_time.average"))

Any operand also accepts a Python callable fn(entity, model), and Call(fn=...) runs Python code in the flow.

Expressions

Expressions use SIMAN syntax, e.g. TRIA(1, 2, 4), NQ(Triage.Queue) > 3 .AND. Size == 2, TNOW - ArrTime. Every expression field has an ƒx Expression Builder (also on right-click, like Arena's Build Expression): pick a distribution, a variable, attribute, resource, queue, statistic or function from lists of the model's own names and fill in its parameters; DISC and CONT are built from a table of values and percentages, cumulative probabilities included. Named expressions go in the Expression data module (the replacement for SIMAN's obsolete DISTRIBUTIONS/DIST).

Group	Supported
Distributions	EXPO UNIF TRIA NORM LOGN GAMM ERLA BETA WEIB POIS DISC CONT JOHN (and the two-letter forms EX UN TR …), optional stream argument
Math	ABS SQRT EXP LN LOG SIN COS TAN ASIN ACOS ATAN HSIN HCOS HTAN INT NINT AINT ANINT MOD AMOD MN MX AMIN1 AMAX1
Time and run	TNOW TFIN NREP MREP J
Queues	NQ AQUE ENTATRANK
Resources and sets	NR MR IRF STATE RESUTIL RESSEIZES MEMBER MEMIDX NUMMEM
Statistics	TAVG TMAX TMIN TNUM TSTD THALF DAVG DMAX DMIN NC
Entities	Entity.Type Entity.SerialNumber Entity.Picture Entity.CreateTime IDENT EntitiesIn EntitiesOut EntitiesWIP NE NG NSG NB NSTO
Schedules

In the browser

python -m simcore.web.build dist          # static site: editor + engine (Pyodide) + example models
python -m http.server -d dist 8000        # try it at http://localhost:8000/

dist/ runs entirely in the browser: the engine is this package running in Pyodide inside a Web Worker, loaded from cdn.jsdelivr.net on the first visit and then cached by sw.js for offline use. Put the folder on any static host — GitHub Pages (.nojekyll included), Netlify or Cloudflare Pages (_headers included):

npx wrangler pages deploy dist --project-name <your-project>   # Cloudflare Pages

Saved models stay in each browser; File → Export JSON gives students a file to hand in. The End button keeps partial results when the page is cross-origin isolated (_headers on Netlify/Cloudflare, sw.js elsewhere). Long runs are slower in the browser than in CPython; a progress panel shows how far a run is.

Random numbers

model.rng_mode = "arena" (default) uses L'Ecuyer's MRG32k3a streams and substreams (legacy 16807 streams when SEEDS are given) with fast classic samplers; "exact" uses statistically exact samplers on the same streams. model.disable_randomness = True replaces every distribution by its mean.

Project layout

Path	Contents
simcore/core.py	entities, queues, resources, variables, simulation loop
simcore/blocks.py	SIMAN blocks
simcore/model.py	elements, expression resolver, replication runner, statistics collection
simcore/transfer.py, guided.py	free-path transporters, conveyors, guided vehicles
simcore/continuous.py	levels, rates, CSTATS, DETECT
simcore/rng.py, stats.py	random numbers and distributions, tallies, half widths
simcore/loader.py	YAML format and Arena-style modules (process, decide, batch, …)
simcore/web/	editor: api.py (shared by server and browser), server.py, build.py (static site), static/ (page, worker.js, browser.js, learn.js, report.js, scenarios.js, analyzer.js, plots.js)
branding/	logo (SVG and PNG)

Tests

python -m unittest discover -s tests       # engine, loader, editor graph conversion, build
# browser tests (headless Chrome; set PLAYWRIGHT_CORE to playwright-core's index.mjs if needed),
# with `python -m simcore.web --no-browser` running:
node tests/ui/editor.test.mjs
node tests/ui/learn.test.mjs
node tests/ui/tutorials.test.mjs
node tests/ui/tools.test.mjs
# the static build across several visits (serve dist/ first):
BASE=http://127.0.0.1:8790/ node tests/ui/revisit.test.mjs

Bahasa Indonesia

Atlverse Simulation adalah aplikasi simulasi sistem diskrit untuk perkuliahan. Mahasiswa cukup membuka https://simulation.atlverse.xyz di browser — tidak perlu instalasi; simulasi berjalan di laptop masing-masing. Fitur untuk kelas:

• Tutorial terpandu (🎓): model antrean pertama, Decide, Assign dan Schedule; setiap langkah diperiksa otomatis.
• Help & istilah (F1): sekitar 70 istilah laporan dan modul dijelaskan dalam Bahasa Indonesia; judul kolom di Summary bisa diklik.
• Cara membaca hasil: penjelasan otomatis di atas Summary (utilisasi, waktu tunggu, Hukum Little, half width).
• Identitas mahasiswa (👤) dan File → Laporan HTML/PDF dengan watermark nama, NIM dan kode verifikasi; dosen memeriksa keaslian laporan lewat File → Periksa laporan.
• Expression Builder (tombol ƒx atau klik kanan di kolom expression): pilih distribusi, variable, resource, antrean atau fungsi dari daftar; DISC/CONT cukup diisi tabel nilai dan peluang (%).
• Tools → Scenarios untuk membandingkan skenario, Tools → Input Analyzer untuk mencari distribusi dari data.

License and trademarks

MIT © 2026 Andrew Tanny Liem.

Atlverse Simulation is an independent project. It is not affiliated with, endorsed by or sponsored by Rockwell Automation. Arena and SIMAN are trademarks of Rockwell Automation, used here only to describe compatibility. The example models are built from the exercises in Altiok & Melamed, Simulation Modeling and Analysis with Arena (Academic Press, 2007).