Implementation of tree pruning method for power transformer design optimization

Abstract

Different optimization algorithms and methods have been proposed in recent decades, most of which are somehow inspired from nature and natural events, especially those that can simply be analyzed and formulated. This work aims to propose a new optimization method inspired from growth and fecundity process of plants, which enable it to prune some branches of the plant or tree, that is, bad parts of the search space which do not meet the problem constrains and rapidly search the whole search space. By analogy, we can prune the search space of an optimization problem, cutting and ignoring its bad parts, the parts that certainly do not meet the problem constrains, so that the search speed and quality can be improved considerably. By improving the search quality, we mean to completely get rid of being trapped in local minimum. There are subtle similarities between pruning a tree to improve its fertility and pruning a search space of an optimization problem to improve the prosperity of optimization algorithm. JMAG-Designer software is used for the finite element analysis.