ECOLOPES Computational Model

Abstract

This deliverable describes the development of the ECOLOPES Computational Model for the ontology-aided generative computational design process, which comprises the translational and the generative process. The translational process involves preparation of datasets that serve as input for the generative process. The generative process involves generating (1) design output concerning spatial organisation and (2) geometric articulation for each design output. Stage 1 of the deliverable focuses on the development of an Answer Set Programming (ASP) approach for the generative process. Stage 2 comprises preparation for utilising additional classes of algorithms (e.g., ASP approach for the translational process and extending the generative process with a Genetic Algorithm and a Machine Learning algorithm. While stage 2 will not reach the required TRL it will set out a clear path for future development of the ontology-aided generative computational design process. Furthermore, this deliverable describes the approach to the validation of the ontology-aided generative computational design process. This includes technical functionality of (1) the key components of the ontology-aided generative computational design process, the ECOLOPES Information Model (EIM) Ontology (D4.2 Interim EIM Ontology), the ECOLOPES Voxel Model (D5.3 ECOLOPES Voxel Model) and the ECOLOPES Computational Model, as well as (2) their various interactions. Secondly, this includes validation of the soundness of the design output of the various stages of the design process. The Vienna Case Study outlined in this deliverable is currently in preparation and will serve this purpose. Finally, given the aim of the ECOLOPES research project to develop a design approach and computational design workflow for use in architectural practice, it is necessary to validate the robustness and usefulness of the approach in a simulated practice context. We operate on the understanding that given a reasonably compact amount of training, an architect with a first degree (BA) in architecture should be able to utilise the conceptual, methodological, and computational aspects of the ontology-aided generative computational design process. For this reason, we use master-level design studios and master thesis projects at TU Wien as testbeds for this purpose. This serves to (1) establish the training that is necessary to enable master-level students to work with the Ecolopes approach, and (2) evaluate design outcomes of Ecolopes projects undertaken by the students.