Electronically Scanned Array (ESA) Performance Prediction Model

In Phase I, we propose to develop novel electromagnetic software that can model the complex requirements of modern electronically scanned array (ESA). We will focus on fast algorithms for rapid analyses of ESA using parallel computers, allowing the efficient analyses of multimode character of an ESA. The efficient algorithm will be achieved by using thin-dielectric sheet for perfect electric conductor and surface integral equation for efficient modeling. The method of moments (MOM) is used to obtain the matrix equation from the integral equation. Then the Huygens equivalence principle is used to seek reduced representation of an antenna element, which can be reused in an array, greatly reducing the memory and CPU requirements. The multilevel fast multipole algorithm will be used to accelerate the solution of the matrix equation that follows from MOM. Then a message passing interface (MPI) driven parallel algorithm will be used to parallelize MLFMA on a Linux cluster. In Phase II, we propose to use the results of Phase I to couple the rapid solution with jumping gene genetic algorithm for the performance optimization of the ESA. Furthermore, we will use fast near-field to far-field transform to obtain the radiation patterns of the ESA rapidly.