Application Of Self Organization Neural Network Technique Essays

1393 Words Feb 4th, 2015 6 Pages
Title: Application of Self-Organization Neural Network Technique (SOM) to Optimize Finite- Element Partial Differential Equation (PDEs) Results in Square-Shaped Structures Analysis.

The finite-element method (FEM) is a computationally method for solving partial differential equations (PDEs) with specific boundary conditions over a domain. When we applying the FEM to a domain, it has to divide to a finite number of elements and nodes. The collections of the elements and nodes form the finite-element mesh, whose quality is an important part in achieving exact numeric result for all finite-element codes.

To achieve a high-quality mesh, we have to decide on:

• The suitable size and topology of the mesh,
• How it should be placed on the area,
• Make decisions regarding the organization of the data on the mesh,

In our research we given the number of nodes on a domain to optimize the productivity of the self-organization neural network method by comparing results from this method by the results that we have achieved by finite- element methods.

On the other hand, A self-organizing map (SOM) is a type of artificial neural network that is trying to find hidden structure in unlabeled data to produce a low-dimensional (two- dimensional), which it discrete representation of map. However, this method was not formulate to optimization problems.

Therefore, some characteristics of a neural network can be summarize to:
• The ability to learning
• Distributed memory
• Fault…

Related Documents