Invisibility cloak designed with global optimizations

Abstract

The biggest hurdle faced in modelling invisibility cloaks using the current methods is the increased scattering caused by the stepwise approximation of the coordinate transformation model. To overcome this, multiple researchers have used Optimization as a means to derive the best parameters for each layer of the invisibility cloak. However, the efficacy of these optimization methods on cloaks of any arbitrary dimensions and number of layers are not known, as most of the current research usually give few examples of where the optimizations have managed to derive the parameters for the lowest scattering possible. Therefore, this project aims to find out the efficacy of this optimization procedure on cloaks of various number of layers and various dimensions, though constructing models, in a bid to test the consistency of the optimized results. Through the research conducted in this project, it was realised that the current optimization method does not achieve the global minimum, which is the lowest result that a mathematical function can give, for multi-layered cloaks of most dimensions. To further reduce scattering in a multi-layered cloak, a method for global optimization and a model to derive the best start points for a solver was developed.