DesignLab
  • Departamento de Engenharia Mecatrônica
    Escola Politécnica - Universidade de São Paulo

Using Hierarchical Models for Requirement Analysis of Real World Problems in Automated Planning

Basbaum, Rosi and Silva, Javier Martinez and Silva, Reinaldo

Journal 
 
Autor 
Basbaum, Rosi and Silva, Javier Martinez and Silva, Reinaldo 
Book Title 
Proceedings of ICAPS 2016. 
Volume 
Página(s) 
101--108 
DOI 
 
Published 
2016 
Month 
Tipo de Documento 
Conference Proceedings 
Abstract 
In the intelligent design field, the early phase of requirement analysis plays a fundamental role, especially when dealing with problems to which an analytic formal solutions is not applied. Automated planning appears in that category - particularly when the target are ”real world” systems. Requirement analysis is exactly where the Engineering Knowledge embedded in the problem is explored to provide clues to the solution. A great effort has been made today in the area of Artificial Intelligence to define a reliable design process for automated planning that includes a Knowledge Engineering early phase. This paper intents to propose a requirements analysis formal procedure that starts by taking requirements for planning problems represented in UML and proceed to an analysis process based on Petri Nets. In fact a similar approach were insert in a tool called itSIMPLE (by one of the authors) using an old version of UML (1.4). As we will see in this proposal the process were not completely formal, even if practical. In the current work an unified Petri Net is fully complied with the ISO/IEC 15.909 standard replace the graph analysis. Using this net we can introduce a formal property analysis including invariant analysis and exploring abstraction from a hierarchical extension of classic Petri Nets. Case Studies are presented, with classic problems from the manufacturing and petroleum industries, aiming to show the differences between the early proposal and the current proposal. 
Keywords 
 
Publisher 
Palo Alto: AAAI, 2016 
URL 
http://icaps16.icaps-conference.org/proceedings/spark16.pdf