Automated assessment of architectural spatial layout configurations using fuzzy logic

Abstract

PurposeConventional approaches of evaluating spatial layout configurations typically involved universal understandings of aspects like connectivity, proximity and visibility, while possibly discarding both partially true solutions and ranges of parameters affecting detailed spatial relations. With the growing need to address spatial uncertainty and ambiguity, the incorporation of methods that embrace soft qualities in design is becoming increasingly significant in spatial layout planning.Design/methodology/approachThe authors introduce a fuzzy-based approach for the automated assessment of architectural spatial layout configurations while addressing ambiguity in layout design. The authors evaluate soft interdependent design qualities like connectedness, enclosure and spaciousness to satisfy multiple mutually inclusive criteria and account for all logical solutions without discarding likely or less likely solutions. The authors analyze spatial entities, parameters and relations and identify rulesets for logical configurations using linguistic variables, fuzzy sets, membership functions and descriptive rule blocks. As a case study, the authors use grasshopper and fuzzyTECH to represent four pilot layout alternates with varying attributes and a case study focusing on one specific spatial criterion.FindingsMultiple complex and nuanced spatial relations were inferred by evaluating spatial outputs and their inherent discrepancies and correlations, thus confirming the assumption that fuzzy-based systems could potentially satisfy multiple mutually inclusive criteria and account for exhaustive logical solutions without discarding preferable, likely or less likely solutions.Originality/valueMost precedent approaches focus on spatial layout design from an occupancy-centered perspective, where occupancy patterns and possibilities are identified in loosely defined spaces or behavioral usage patterns. The added value in this paper involves including a wide array of spatial inputs to describe soft spatial qualities using nuanced rule-based descriptors.