Performance based design evolution : the use of genetic algorithms and CFD

This paper describes a performance-based evolution model using Genetic Algorith as the evolution algorithm and CFD as the evaluation mechanism. The advantages of such an evolutionary performance-based design approach is that diverse instances of the state space can be investigated in relation to specific goal requirements that will enhance the possibility of discovering a variety of potential solutions. The model allows the user to explore and visualize the design evolution and its form generation in an attempt to stimulate the designer creativity that might contribute to their output. The process uses an iterative approach that allows design to be evaluated using CFD analysis automatically to maximize several thermal and ventilation criteria. Design change will then be performed, re-meshed and displayed based on the evolutionary algorithms. The process allows the user to experience the morphing of the design based on its performance and continues until the designer has the opportunity to visualize the evolution of the final set of design alternatives.