Modelling the dynamics of complex early design processes: an agent-based approach. (30th October 2017)

• Record Type: Journal Article
• Authors: Fernandes, João Ventura; Henriques, Elsa; Silva, Arlindo; Pimentel, César
• Is Part Of: Design science. Volume 3(2017)
• Publication Date: 2017-10-30
• Subjects: Complex design, -- early design phases, -- process modelling, -- agent-based simulation, -- early design planning; Design -- Research -- Periodicals; New products -- Management -- Periodicals; Design; Design -- Research; Electronic journals; Periodicals; 658.5752
• DOI: 10.1017/dsj.2017.17
• Languages: English
• ISSNs: 2053-4701

Abstract

Abstract : Among the different phases of complex design processes, early design is the most dynamic and unpredictable stage since it involves a great deal of uncertainty, concurrency of activity streams, collaborative design iterations, and distributed and adaptive decision-making behaviour in response to both organizational commitments and to the occurrence of unforeseen events. This paper argues that current activity-based modelling approaches have limited ability to capture the dynamics of complex early design processes and explores novel modelling approaches. The development of an Agent Model for Planning and rEsearch of eaRly dEsign (AMPERE) aiming to capture various facets of uncertainty, iteration, collaboration and adaptation is described. The model was developed to tackle early design phases of complex systems, with the ability to deal with changes in requirements coming in and affecting the subsequent design evolution while design tasks are on-going. Initial results from agent-based simulations are presented, showing how the agent-based approach can support industrial organizations evaluating likely early design project performance and understanding complex cause–effect relationships that may affect project outcomes. Early design planning support from the agent model is demonstrated through an investigation to the likely project performance for varying levels of externally driven requirements change.