The tool Production Simulation and Analyses is a simulation library, which enables the user to simulate and evaluate an arbitrarily configurable production area. The focus of the simulation library is on production planning and control (PPC) and in particular the tasks: Throughput planning, order release, capacity control and dispatching. The model contains a number of well-known PPC heuristics for these tasks, each of which can be freely parameterized.
In addition to the flow simulation, which is realized as a discrete event simulation using the Python framework SimPy (https://simpy.readthedocs.io/en/latest/), the simulation library offers the user a variety of possible data exports and integrated analyses, which are provided using Matplotlib (https://matplotlib.org/) and seaborn (https://seaborn.pydata.org/).
Futhermore, the model includes an experimental layer for parallel and sequential simulation using SLURM Workload Manager (https://slurm.schedmd.com/).
Production Simulation and Analyses is completely based on Python. The packages needed for the simulation are documented in requirements.txt.
To give an overview, the required packages are listed here as well.
| Package | Version |
|---|---|
numpy |
1.19.1 |
pandas |
1.1.0 |
simpy |
4.0.1 |
scipy |
1.6.0 |
matplotlib |
3.6.0 |
seaborn |
0.12.1 |
The simulation library is mainly configured via the classes ModelPanel and PolicyPanel, which are included in control_panel.py. While ModelPanel contains, in particular, more basic structural configuration decisions of the production area to be considered, PolicyPanel contains the concrete configuration of the individual PPC tasks. In dependence on the use of the so-called batch_manager, the settings which can be varied can be specified by a parameter dictionary, which is iterated by the exp_batch_manager.py and provided by exp_parameters.py.
The simulation library runs automatically if a user starts the exp_batch_manager.py and defines the upper and lower experiment limits. During the run, various modules are loaded, and data is generated and processed. For the data evaluation, the setting in exp_manager.py is to be considered in particular. Here it can be set which exports/analyses are to be carried out by the simulation library.
For further information, please refer to the annotation of the code in the individual files. Furthermore the author thanks Arno Kasper for providing the tool Process Sim (https://github.com/ArnoKasper/ProcessSim), which is the basis for the created simulation library, and for the permission for its further usage.