Skip to content

Basic manufacturing system

Goal: build a small manufacturing system, run jobs through it, then inspect basic performance metrics.

1) Build a small shop floor with two servers

from simulatte.environment import Environment
from simulatte.job import ProductionJob
from simulatte.server import Server
from simulatte.shopfloor import ShopFloor

env = Environment()
shopfloor = ShopFloor(env=env)

s1 = Server(env=env, capacity=1, shopfloor=shopfloor)
s2 = Server(env=env, capacity=1, shopfloor=shopfloor)

2) Create jobs with a routing

Each job has:

  • servers: the routing (operation order)
  • processing_times: time per operation
  • due_date: absolute simulation time
jobs = [
    ProductionJob(env=env, sku="A", servers=[s1, s2], processing_times=[5.0, 3.0], due_date=40.0),
    ProductionJob(env=env, sku="A", servers=[s1, s2], processing_times=[4.0, 2.0], due_date=40.0),
    ProductionJob(env=env, sku="B", servers=[s1, s2], processing_times=[6.0, 2.0], due_date=40.0),
]
for job in jobs:
    shopfloor.add(job)

3) Run and read results

env.run()

print(f"Jobs completed: {len(shopfloor.jobs_done)}")
print(f"Avg time in system: {shopfloor.average_time_in_system:.2f}")
print(f"s1 utilization: {s1.utilization_rate:.1%} (avg queue len: {s1.average_queue_length:.2f})")
print(f"s2 utilization: {s2.utilization_rate:.1%} (avg queue len: {s2.average_queue_length:.2f})")

tardy = sum(1 for j in shopfloor.jobs_done if j.late)
print(f"Tardy jobs: {tardy}/{len(shopfloor.jobs_done)}")

Notes

  • Simulatte avoids global singletons: pass the same env to every component you want in the same simulation.
  • Every server, shop floor, pre-shop pool and router has an id: the name= you pass, or a generated server-0, server-1, ... in creation order. Jobs are numbered job-0, job-1, ... per environment. The build_*_system builders name their entities wc-0, wc-1, ..., shopfloor, router and psp; pass prefix= to build several systems in one environment.

Next