Graph-Based Problem-Solving and Representation: Levels of Deployment in Computational Design Process
Dragana Ćirić
https://doi.org/10.60152/7clkf2xv
Abstract: The paper addresses the question of the deployment of ‘dependency graph’-based problem-solving and representation methods. The question is observed in terms of decomposition levels of the computational design process and the subject of computation, or graph representation and modelling. Based on the stated criteria that firstly distinguish process graphs from object-based or formally congruent graphs, and secondly identify sublevels within the formed classes, the paper explains 1. two levels of the first group (1a. the level of a complete design problem-solving algorithm (metalevel process propagation graphs) and 1b. the level of smaller problem-solving clusters of methods (specific operations graphs, or sequential process propagation graphs)) and 2. specificities of the second group at the level of content that is to be modelled as a graph (subject-based graphs, including formal network-graph congruency). The importance of determining deployment levels has been recognised in the need to clarify, in a systematised way, the graphs’ use, pursuing elaboration of each application level’s scope or problem-solving coverage, and complexity considering the subject of analysis and representation. The study supports the aim of developing new and optimised process workflows and better-informed conduct when approaching computational design problems and tasks, leading to a higher standard of performance at each scale and construction/design stage of the computational problem-solving path and representation. In that respect, the stated arguments are supported by the examples drawn from the experimental case of dependency (network and navigation) graph application in the field of architectural and urban computing, defined and tested in Grasshopper. The structure and results of the thereby created graphs, based on the investigated subjects of transportation infrastructure design and analytics, dynamic localisation, and path-finding, are parsed in a way that best corresponds to the intended instructional explanation. The case study provides supporting evidence in the form of graph-based workflows and formal geometric outputs resulting from their propagation – i.e., it supplements arguments diagrammatically and provides illustrations of the stated points.
Keywords: Network Graphs, Diagrams, Design Process, Computational Design Methodology, Computational Problem-Solving, Algorithmic Thinking, Urban and Architectural Computation, Smart Cities, Intelligent Cities, Intelligent Architecture
How to cite this Paper (Harvard referencing style):
Ćirić, D. (2024) ‘Graph-Based Problem-Solving and Representation: Levels of Deployment in Computational Design Process’, in R. Bogdanović (ed.) On Architecture — Shaping the City through Architecture, Proceedings. Belgrade, Serbia: STRAND, pp. 171–181.