You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Bernhard Wally 54146d9aa4
The one and only commit.
2 years ago
iec62264 The one and only commit. 2 years ago
metamodels The one and only commit. 2 years ago
move-shuttle The one and only commit. 2 years ago
pddl The one and only commit. 2 years ago
results The one and only commit. 2 years ago
transformations The one and only commit. 2 years ago
README.MD The one and only commit. 2 years ago

README.MD

Support Material for “Leveraging Iterative Plan Refinement for Reactive Smart Manufacturing Systems”

This repository contains supplemental resources for the T-ASE paper “Leveraging Iterative Plan Refinement for Reactive Smart Manufacturing Systems”. Specifically, it provides detailed information about the mapping from IEC 62264 to PDDL:

Creating PDDL actions from IEC 62264 process segments

Excerpts from an IEC 62264 model that are relevant to generate a PDDL action from an IEC 62264 process segment. The example files are provided in the sub-folder move-shuttle.

Mapping Rules

A detailed guide to the mapping rules between IEC 62264 and PDDL that have been applied in the context of this paper are explained in the README.MD in the sub-folder transformations.

IEC 62264 Models

Full IEC 62264 model of the production system, as it is used in the paper, for the test-run with one product to be manufactured. The core IEC 62264 model that is used to generate the PDDL domain description as well as most parts of the PDDL problem file is provided in iec62264/CIIRC-Testbed-TASE-1.iec62264. The IEC 62264 model that describes the envisioned goal state of the production system and is used to generate the goal clause of the PDDL problem file is given in iec62264/CIIRC-Testbed-TASE-Goal-1.iec62264.

PDDL Files

The generated PDDL files are provided in sub-folder pddl. The domain file is used together with the problem description to compute a plan. The plan file itself is already a computed artifact returned from the PDDL solver, which in this use case was FastDownward. The following call can be used to re-create the plan file (potentially with a different sequence of actions).

fast-downward.py --plan-file plan-CIIRC-Testbed-TASE-1.pddl domain-CIIRC-Testbed-TASE.pddl problem-CIIRC-Testbed-TASE-1.pddl --evaluator h1=ff --search eager_greedy([g,goalcount,cg,h1],[h1])

Metamodels

Full metamodels for IEC 62264 and for PDDL, as they have been used in the context of this paper, are provided in sub-folder metamodels. Both are defined using Ecore as the meta-metamodel.

Results

Results of all evaluation runs (as tab-separated values) are provided in sub-folder results. Nomenclature of the files is as follows: out-<numOfProducts>-<eagerInitial>-<eagerForecast>-<parallel>.csv. Where numOfProducts is the number of products to be produced in one production run (from 1 to 7). eagerInitial states that the initial plan was searched for using an eager search algorithm (true) or a lazy one (false). eagerForecast is true if the forecasting service was realized using an eager search algorithm, otherwise false. parallel is true if the found plans are being parallelized a-posteriori, otherwise false.