Factory Domain - version 1.0

Learning in Planning with Temporally Extended Goals and Uncontrollable Events




Introduction

Specification

Format

Data Files

Extensions

Data Files

We developed an automated procedure to randomly generate the data set. In this procedure, each instance is created and identified by three numbers:
  1. Number of repositories.
  2. Number of machines.
  3. Seed for random generator.
Given the number of repositories and machines, the procedure will randomly determine the initial system state, the possible actions/events and the LTL formula that must be satisfied. Two instances generated with the same number of repositories, machines and seed are guaranteed to be exactly equal.

The procedure was implemented in C++ and the source can be downloaded from the link below. You can generate the exactly instances used in the paper submitted to ECAI by running the script generate_ECAI.sh, included in the tar file.

factory-benchmark-generator.tar.gz.

For additional information regarding solution generation, execute the binary generate after compilation.

Back to the beginning.
l>