WP3: AI-Assisted Modeling and Optimization | CLEM - Combining Live and Exploratory Modeling

WP3 investigates the integration of artificial intelligence and machine learning techniques to provide intelligent assistance for model creation, validation, and optimization.

Objectives

This WP develops comprehensive language workbench support for live and exploratory modeling. It combines execution engine integration (from WP1) with user-facing interfaces (from WP2), resulting in a practical platform for xDSML developers. The workbench provides end-to-end support for designing, executing, and exploring models live.

Partner Contributions

U.Rennes: Leads the integration of LEM capabilities into the language workbench (GEMOC Studio), bringing expertise in model execution platforms, xDSML design, and workbench architecture.

JKU: Contributes expertise in exploratory modeling, test case generation via search-based techniques (MOMoT), and comprehensive testing/documentation practices to ensure workbench quality and usability.

Tasks

Task T3.1: Workbench Support for xDSML Design with LEM (Lead: U.Rennes)

Description: Integration of live modeling and exploratory modeling capabilities into the GEMOC Studio language workbench. This task focuses on providing seamless support for designers to leverage LEM features during xDSML development and usage, including execution control panels, interactive state inspection, and real-time debugging features.

Challenge: Maintaining workbench responsiveness and usability while integrating complex LEM features without overwhelming users.

Task T3.2: Generative Approach Integration (Lead: U.Rennes)

Description: Integration of search-based and generative techniques for model generation and exploration within the workbench. This includes automated test case generation via search algorithms, guided exploration through optimization heuristics, and systematic exploration of design alternatives to accelerate model development and validation.

Challenge: Ensuring search algorithms scale to realistic model sizes and deliver actionable results.

Task T3.3: Workbench Release, Testing, and Documentation (Lead: JKU)

Description: Comprehensive testing, quality assurance, and documentation of the LEM-enabled workbench. This includes integration testing, performance benchmarking, user testing with case studies, and creation of tutorials, guides, and API documentation for developers adopting the workbench.