A GPU parallelization for grid generation of fuzzy topographic topological mapping

Abstract

The fundamental in grid generation for multidimensional geomet-rical shape is to construct its grid form. Techniques for creating the grid forms and the smaller shapes formed is the basis of grid generation [1]. The form it-self will determine the quality of the generation process and well grid-construc-ted to describe numerical evaluation and speed. In this paper, a structured grid for Fuzzy Topographic Topological Mapping (FTTM) is proposed to improve the quality of grid in numerical perspectives. FTTM is a mathematical model to de-tect the neuro-inverse magnetic region for neurological disorder [2]. The detec-tion region is based on 4 vertices of FTTM and homeomorphic to each other [3]. A computable homeomorphism will use to define the vertices and edges compo-nents of FTTM. The edges represent their homeomorphisms. A topology on grid generation of FTTM addresses the fuzzy topographic and mapping algorithm. The mathematical modeling of FTTM performs the grid structure, design the grid-connected and synchronize the grid generation. For large and extended FTTM, mesh refinement coupled with fine granularity is used to generate the grid via multi-component and multi-version parallelization scheme. The detail of the construction and performance of the strategy is elaborated, evaluated and reported in the paper.