Design synthesis for Multi-X : a 'life-cycle consequence knowledge' approach
Product design decisions can result in unintended consequences that propagate across multiple life-phases such as manufacturing, use and disposal. If designers are to generate 'life-oriented' solutions, handling this phenomena is a necessity. Due to the sequence of life-phases, knowledge of such 'life-cycle consequences' (LCCs) is generated late, after decisions have been committed. Thus, designers have difficulties in foreseeing LCCs co-evolving with their solution. Further, a literature review established that, designers currently lack adequate support to foresee and explore LCCs during synthesis. To address this 'Design Synthesis for Multi-X (DsF∑X)' problem, this thesis proposes, implements and evaluates a computational 'Knowledge of life-cycle Consequences (KC)' approach. The establishment of a phenomena model disclosing how LCCs are generated from two different conditions has highlighted the necessity of concurrent 'artefact' and 'Iifephase system' synthesis. This provided a foundation of how to model and timely utilize LCC knowledge for revealing LCCs co-evolving with a solution description. This resulted in a framework for the 'KC' approach consisting of: the 'LGG knowledge modelling frame' which presents a formalism of 'what' elements to acquire and model for an application domain, together with how to structure the established relationships into 'LCC inference' and 'LCC action' knowledge; an 'artefact life modelling' frame which provides a formalism for describing 'artefact life' compositional models that support the inference of LCCs; and the 'operational frame' which discloses principles of how a LCC knowledge model can be utilized to amplify the human designer's capabilities. By identifying system requirements, an architecture and knowledge codification schemes, the framework was realized as a Knowledge Intensive GAD prototype, 'FORESEE', for the thermoplastic component domain. An evaluation of FORESEE established that the 'KC' approach integrates synthesis with foreseeing multiple LCCs. This is fundamentally different from first generating a candidate solution and afterwards analysing the solution for conflicts with artefact life issues. The 'KC' approach thus provides a step towards realizing pro-active DsF∑X support. However, further work is required to the framework and FORESEE to practically exploit its utilization.