psplib 0.2.0

A parser for project scheduling instances.

PSPLIB

This library implements parsers for various project scheduling benchmark instances, including:

• Resource-Constrained Project Scheduling Problem (RCPSP)
• Multi-Mode Resource-Constrained Project Scheduling Problem (MMRCPSP)
• Resource-Constrained Project Scheduling Problem with Minimal and Maximal Time Lags (RCPSP/max)
• Resource-Constrained Multi Project Scheduling Problem (RCMPSP)

psplib has no dependencies and can be installed in the usual way:

pip install psplib

Example usage

>>> from psplib import parse
>>> instance = parse("data/j301_1.sm", instance_format="psplib") 
>>> instance.num_resources
4

>>> instance.resources
[Resource(capacity=12, renewable=True), ..., Resource(capacity=12, renewable=True)]

>>> instance.num_activities
32

>>> instance.activities
[Activity(modes=[Mode(duration=0, demands=[0, 0, 0, 0])], successors=[1, 2, 3], delays=None, name=''), 
 Activity(modes=[Mode(duration=8, demands=[4, 0, 0, 0])], successors=[5, 10, 14], delays=None, name=''),
 ...,
 Activity(modes=[Mode(duration=0, demands=[0, 0, 0, 0])], successors=[], delays=None, name='')]

See this notebook for a slightly longer example demonstration of PSPLIB.

Instance formats

psplib implements parsers for commonly used project scheduling instance formats, listed below. To parse a specific instance format, set the instance_format argument in parse.

1. psplib: The PSPLIB format is used by the PSPLIB library to describe RCPSP and MMRCPSP instances.
2. patterson: The Patterson format: used for RCPSP instances, mostly used by the OR&S library. See this website for more details.
3. rcpsp_max: The RCPSP/max format is used for RCPSP/max instances from TU Clausthal.
4. mplib: The MPLIB format is used for RCMPSP instances from the MPLIB library.

Instance databases

The following websites host widely-used project scheduling benchmark instances.

• PSPLIB contains different problem sets for various types of resource constrained project scheduling problems as well as optimal and heuristic solutions.

• OR&S project database is the research data website of the Operations Research and Scheduling (OR&S) Research group of the Faculty of Economics and Business Administration at Ghent University (Belgium). OR&S is very active in the field of project scheduling and has published instances for many project scheduling variants.

• TU Clausthal provides RCPSP/max benchmark instances.