Getting started with SimPM

Installation

Install SimPM from PyPI:

pip install simpm

A minimal example

This minimal model represents a single project activity that requires one crew. The activity waits until the crew is available, performs the work, and then releases the resource. The SimPM engine models activities with do (time spent working) and get / put (queueing for and releasing resources), rather than raw timeout calls.

This example shows how to:

import simpm
from simpm.des import Environment, Resource, Entity
from simpm.dist import norm

# 1) Create the simulation environment
env = Environment("Single activity")

# 2) Define resources (e.g., one crew)
crew = Resource(env, name="Crew", capacity=1)

# 3) Create an entity to perform the activity
activity = Entity(env, "Activity", print_actions=False, log=True)

# 4) Define the activity process
def activity_process(entity, resource):
    # Request one unit of the crew resource and wait until it is available
    yield entity.get(resource, 1)

    # Perform the work (duration drawn from a normal distribution)
    # Note: Entity.do accepts either a numeric duration or a simpm.dist distribution.
    yield entity.do("work", norm(10, 2))

    # Release the crew
    yield entity.put(resource, 1)

    print(f"{entity.env.now:.2f}: {entity} finished.")

env.process(activity_process(activity, crew))

# 5) Run the simulation
simpm.run(env)

print(f"Project finished at t={env.now:.2f}")
print(activity.schedule())

Why SimPM instead of SimPy?

While SimPM is inspired by SimPy’s process-based discrete-event simulation approach, it is purpose-built for project and construction management with features that SimPy requires extensive custom code to achieve:

Entities with Attributes

SimPy entities are minimal; you must build your own classes and manage attributes manually. SimPM’s simpm.des.Entity comes with built-in support for scheduling, waiting times, and custom attributes—tracking entity state is effortless.

Automatic Logging and Analytics

SimPy does not log events or metrics automatically; you write custom instrumentation for each model. SimPM automatically tracks events, queue lengths, waiting times, and resource utilization, storing results in pandas-friendly tables ready for analysis. No extra code needed.

Integrated Distributions

SimPy offers no built-in distributions; you manually sample from NumPy or SciPy. SimPM includes project-friendly distributions—triangular, beta, trapezoidal, normal, exponential, uniform, and empirical—all integrated with Entity.do() for seamless stochastic simulation. Monte Carlo schedule risk analysis becomes straightforward.

Dashboard and Visualization

SimPy leaves visualization entirely to you. SimPM includes an optional interactive dashboard that automatically displays timelines, queue behavior, bottlenecks, and completion time distributions after each run.

Waiting Times and Utilization

SimPy requires you to instrument your model to extract waiting times and utilization metrics. SimPM calculates and reports these automatically—no extra instrumentation code required.

Built for Project Management

SimPM’s design focuses on the questions project managers ask:

  • How long will this project take (with uncertainty)?

  • Where are the bottlenecks in crews, equipment, or workspaces?

  • What happens if we change crew size or equipment plans?

All the tools for answering these questions—distributions, logging, dashboards, and entity tracking— are integrated and ready to use.

Next steps