Reconstruction of a Layered Dielectric Cylinder Using a Split Particle Swarm Optimization

Reconstruction of a layered dielectric cylinder is formulated with cost functions, which are minimized by a split PSO. In the split PSO, particles are split into, for example, two groups. Particles of a group evaluate a cost function and those of the other group do another cost function. Particles change their positions according to experience of their own and both groups to search the common minimum point of both cost functions. Numerical examples show that the split PSO saves computational cost and tends to reduce traps at local minimum points.