Penalty Function Solution to Pattern Syn- Thesis of Antenna Array by a Descent Algo- Rithm

In this paper, an algorithm based on penalty cost function for synthesizing flat-top patterns is proposed. A descent algorithm (DA) as its optimizing approach is proposed in this paper as well. Apparently, whole algorithm efficiency totally depends on the DA. Unlike traditional descent method, the DA defines step length by solving a inequality, instead of Wolf or Armijo-type search rule, stimulation results indicate that it can improve the computational efficiency. Under mild conditions, we prove that the DA has strong convergence properties. Several numerical examples are presented to illustrate the effectiveness of the proposed algorithm. The results indicate that the approach is effective in the pattern shape precisely in both mainlobe and sidelobe region for arbitrary linear arrays.