A Hybrid Approach of Dynamic Programming and Genetic Algorithm for Multi-criteria Optimization on Sustainable Architecture design

ABSTRACT

This paper proposes a method to find heuristic solutions in early decision stages for sustainable building design. This method integrates dynamic programming and genetic algorithm into the design workflow, by separating design criteria into three decision stages to reduce the complexity. In each stage, the designer setups parameters related to sub-criteria and objective functions to optimize by Genetic Algorithm. Dynamic programming is used to guide the search and combine solutions for all stages. An actual project is used to test and demonstrate the proposed method. For implementation, we develop a parametric design system with Rhino, Grasshopper and HLGA to provide a graphical design environment, in which heuristic solutions for each decision stage are found and combined.

GRAPHICAL ABSTRACT

Keywords:

• dynamic programming
• genetic algorithm
• sustainable architecture design

ACKNOWLEDGMENTs

This research was supported by the National Science Council of the ROC, Taiwan for the project number: NSC 102-2221-E-011 -120.