Kun-Tai Kao, Teng-Wen Chang & Ih-Cheng Lai
Learning Construction Decision Making with Arbitrator - Competing and Evolving in Dynamic Role Interplay
Abstract: Among construction process, decision making is often depended on different construction methods invoked within different events occurred. The re-action sequence is hard to predicate and hard to reproduce in different situations, also. For resolving this problem, we adopt a model called Dynamic Role Interplay Process (DRI). Based on the mechanism of DIM (Dynamic Idea Map) (Lai and Chang 2005), competing learning and evolving learning between different roles, and then the knowledge within roles will involve evolving learning. We adopt learning metaphors in design sequence generations. Competing among design events and concepts are treated as the agents who are competing for surviving in the sense of genetic programming. And Roles start to evolve to different knowledge or preserving the core ontology of design for the evolving learning concept. Each process follows the metaphor of natural selection and design problem solving paradigms in design domain. A simulation agent-based system is implemented in CELA.
Full text: content.pdf (1,762,501 bytes) (available to registered users only)
Permission to reproduce these papers has been graciously provided by the Technische Universitšt Dresden.