A Method Based on Interactive Self-Organizing Data Analysis Technique A(ISODATA)for Symmetric Traveling Salesman Problem

Abstract

In this paper, a new algorithm for symmetric Traveling Salesman Problem by ISODATA is proposed. However this is a kind of hierarchical partitioning approach, the solver exhibits high performance of formation that the distribution of cities is uniform. This advantage originates in the knowledge implementation which is well suited to iteration process and high interaction among clusters of ISODATA. Therefore the theoretical analysis of dynamics of ISODATA is very complicated mathematically and the authors exhibit the activity of the system by experiments, for example, effects of parameters, evaluation of test data sets on common library(TSPLIB), uniform data sets, and dynamic data sets.