The construction of rapid algorithms for constructing molecular phylogenetic trees based on the principles of metaheuristic algorithms

点击翻译按钮获取中文摘要

There are two main types of algorithms for constructing molecular phylogenetic trees. The first is those based on the distance matrix, while the other tries to search for an optimal binary tree under minimizing or maximizing given objective functions. The Maximum Parsimony, the Maximum-Liklihood, and the Minimum mean square are the typical examples of the latter Algorithms.Although it is an intractable problem to find an optimal solution in the huge feasible solution space in the second cases, they have not ever realized the importance of the strategy for how to overcome the hugeness of the solution spaces, and have never investigated the methods of local searches. As a result, they neglect the possibility of finding better solutions other than those in the neighborhood of the greedily chosen initial solution.We assert here that the so-called methaheuristic algorithms do work well in the above procedures of local searches. And actually, we have shown from the simulations that Tabu search, one of the well-known methaheuristic algorithms, greatly contribute the expand the traverse of the local search trails.