Conference Agenda

The applicability and scalability of design adaptations utilizing reinforcement learning can be broadened by using graph-based approaches instead of rigid vector- or grid-based ones. However, graph-based approaches often require a high number of simulations to converge. To reduce the simulation effort in the mechanical optimisation, the reinforcement learning setup is enriched with task-specific causal and physically based information. This work systematically investigates the influence of this additional information on the efficiency of design adaptations using a factorial test design.

Implementing novel interaction modalities to CAD systems is raising questions about suitability of types of thinking employed and appropriateness of existing workflows used in CAD in this new context. The study reported in this paper explores the potential for introduction of alternative activities into existing workflows, proposed by designers interacting with 3D shapes using gestural interaction. Findings propose introduction of sculpting and forming paradigms that may reduce the amount of work required to create more complex forms.

Conceptual design methods rarely optimize both requirement fit and cross-principle compatibility, leaving a gap in generating coherent early-stage solutions. Here, we introduce a mixed-integer linear programming formulation that selects one solution principle per function while jointly minimizing local requirement mismatch and system-level incompatibility. Using a small case study, we show how a trade-off parameter controls the balance between functional quality and integration robustness. The results demonstrate that the approach enables transparent, compatibility-aware conceptual synthesis.

Functional decomposition shapes early design decisions but is largely qualitative, leaving units and measures implicit. This work introduces the Quantitative Functional Decomposition Problem, which formalizes functions and interfaces with measurable quantities, making decomposition solvable as a quantified planning problem. Two case studies show that the approach gives immediate feedback on the admissibility of functions and their connections. Design engineers get consistent quantified structures, which speed up iteration, reduce work and set targets for subsequent steps in the design process.