Parallelization of Tree Growth Competition Algorithm and its application in automated design of line antenna

摘要

Aiming at the long runtime in the automated design of tree-shaped line antenna, parallelization of Tree Growth Competition Algorithm (TGCA) is presented based on master-slave prototype. Master process performs the competition model of forest dynamical evolution. Slave process performs concurrently the growth-competition model of natural tree to complete time-consuming sunlight fitness calculation. A novel tree-shaped line antenna is designed with end-fire and high gain. It has 5 branches, gain of 9.048dBi, and bandwidth of 8.45%. And the parallel efficiency reaches to 91% in “Yuanmou-IV” Beowulf cluster system of 10 nodes, which validates the parallelized algorithm.