Open Access Open Access  Restricted Access Subscription Access

Artificial Neural Network Based Hybrid Algorithmic Structure for Solving Linear Programming Problems


Affiliations
1 Department of CSE, Annamalai University, Tamilnadu, India
 

Linear Programming Problems are mathematical models used to represent real life situations in the form of linear objective function and constraints various methods are available to solve linear programming problems. When formulating an LP model, systems analysts and researchers often include all possible constraints although some of them may not be binding at the optimal solution. The presence of redundant constraints does not alter the optimum solution(s), but may consume extra computational effort. Redundant constraints identification methods are applied for reducing computational effort in LP problems. But accuracy of the LP problems goes down due to this reduction of loops and constraints. To achieve optimality in accuracy and also in computational effort, we propose an algorithm, called, hybrid algorithm, it trains the constraint and parameter before applying the formal methodology.

Keywords

Linear Programming, Redundant Constraints, Load Forecasting, Training Parameters.
User
Notifications
Font Size

Abstract Views: 253

PDF Views: 0




  • Artificial Neural Network Based Hybrid Algorithmic Structure for Solving Linear Programming Problems

Abstract Views: 253  |  PDF Views: 0

Authors

L. R. Arvind Babu
Department of CSE, Annamalai University, Tamilnadu, India
B. Palaniappan
Department of CSE, Annamalai University, Tamilnadu, India

Abstract


Linear Programming Problems are mathematical models used to represent real life situations in the form of linear objective function and constraints various methods are available to solve linear programming problems. When formulating an LP model, systems analysts and researchers often include all possible constraints although some of them may not be binding at the optimal solution. The presence of redundant constraints does not alter the optimum solution(s), but may consume extra computational effort. Redundant constraints identification methods are applied for reducing computational effort in LP problems. But accuracy of the LP problems goes down due to this reduction of loops and constraints. To achieve optimality in accuracy and also in computational effort, we propose an algorithm, called, hybrid algorithm, it trains the constraint and parameter before applying the formal methodology.

Keywords


Linear Programming, Redundant Constraints, Load Forecasting, Training Parameters.