Title: From Capability Models to Automated Planning: An AAS-Native Approach for Automatic PDDL Generation

Abstract:

Verifying whether a specific production layout accommodates all necessary manufacturing sequences is a critical task for engineers. While automated planning tools are well-suited to address these queries, the process of encoding these problems into the Planning Domain Definition Language (PDDL) requires specialized knowledge that production engineers generally do not possess. In the context of Industry 4.0, Asset Administration Shells (AAS) have become the standard framework for representing digital twins of industrial assets. This study demonstrates that AAS capability models, when organized according to four key Industry 4.0 standards—VDI 3682 for process descriptions, IEC 61360-1 for semantic property qualification, IDTA 02011 for type hierarchies, and IDTA 02016 for instance descriptions—contain all the requisite data to automatically generate comprehensive PDDL problems.

In contrast to previous methods that relied on PDDL-specific submodels, our methodology derives all planning components from domain-level descriptions of resource functions, known as capabilities. This allows engineers to define capabilities without needing to understand PDDL syntax or planning concepts. We present an extraction algorithm capable of converting distributed Multi-AAS architectures into fully formed PDDL planning problems. To validate this approach, we applied it to AAS models of a laboratory production system, comparing four distinct layout variants through optimal planning. This demonstrates how engineers can systematically evaluate design trade-offs by simply modifying the AAS model and triggering the regeneration of the planning domain.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC