A hybrid real-binary particle swarm optimization (hpso) algorithm: concepts and implementations

Introduction Hybrid real-binary problems have been widely appreciated in electromagnetic (EM) system optimizations. In these problems, a vector to be optimized consists of both real and binary components, which immediately lead to many applications such as material selection, topology exploration, etc. A similar problem, which is sometimes referred as a mixed-integer optimization, deals with the hybridization of real numbers and integers. However, since an integer can be always digitized into 0´s and 1´s, the optimization of binary variables is more essential, and has more flexibilities in practical problems such as optimizing a complex EM structure via the pixel-filling method.