Hongzhi Wang, University of Georgia

  • Abstract: Design of experiments plays important roles in all fields of modern science and engineering. Efficient designs are to be used in order to extract maximum information from the data. However, identifying optimal designs are not necessarily easy for complex real life applications, which are becoming increasingly common in practice. Theoretical results are not widely available for such applications and those that are available only exist for special cases. Several optimization algorithms are used to identify optimal designs, while each algorithm usually targets on one design type only. Here we propose a new nature-inspired evolutionary optimization algorithm which works efficiently on several different types of design problems. Simulation studies establish its superiority over different competing algorithms, in terms of both precision and CPU times.