- Collection:
- Atlanta University and Clark Atlanta University Theses and Dissertations
- Title:
- Training neutral networks using genetic algorithms, 1996
- Creator:
- Collins, Samuel R.
- Date of Original:
- 1996-05-01
- Subject:
- Degrees, Academic
Dissertations, Academic - Location:
- United States, Georgia, Fulton County, Atlanta, 33.749, -84.38798
- Medium:
- theses
- Type:
- Text
- Format:
- application/pdf
- Description:
- Various schemes for combining genetic algorithms and neural networks have been proposed in recent years. Both genetic algorithms and neural networks are computational paradigms that are loosely based on biological concepts. Most neural net learning algorithms require gradient and error information for training. If used for learning, genetic algorithms do not depend on the use of gradient information for training. They are therefore free from the usual constraints that other learning algorithms face. This thesis explores the use of genetic algorithms to train neural networks for pattern recognition. In this research the neural network utilized is a fixed multilayered feed-forward network. The method that is proposed yields both a set of weights and a set of squashing functions for the neural network. The approach evaluates a population of individuals that encode the attributes of the neural network. The squashing function assigned to any neural node is chosen from a master set of available functions. The research suggests that genetic algorithms provide a promising new avenue for training neural networks.
- External Identifiers:
- Metadata URL:
- http://hdl.handle.net/20.500.12322/cau.td:1996_collins_samuel_r
- Rights Holder:
- Clark Atlanta University
- Holding Institution:
- Atlanta University Center Robert W. Woodruff Library
- Rights:
-
