ARTIFICIAL NEURAL NETWORK IMPLEMENTATION FOR SOLVING LINEARPROGRAMMING MODELS

Abstract

This work does not use the classical methods (simplex method, Branch and Bound techniques) which were normally used for solving Linear programming models. The proposed algorithm was considered for implementation with Artificial Neural Network (ANN) using MatLab tool box. It was found that implementation of the neural network will provide comprehensive results when applied with any linear programming models. Besides Artificial Neural Networks are artificial intelligence methods for modeling complex target functions, and are considered to be among the most effective learning methods currently known. Implementation in solving linear programming models became very interesting, as ANNs became appropriate solution where a huge data (number of variables and constraints) is considered. In this work, general model of ANN specified for solving the problem of linear programming will be shown and discussed.The results show a great improvement in prediction of results with a minimum percentage error.