Open AccessON NON-QUADRATIC PENALTY FUNCTION FOR NON-LINEAR PROGRAMMING PROBLEM WITH EQUALITY CONSTRAINTS
Author(s) -
Raju Prajapati,
Om Prakash Dubey,
Ranjit D. Pradhan
Publication year - 2019
Publication title -
international journal of students research in technology and management
Language(s) - English
Resource type - Journals
ISSN - 2321-2543
DOI - 10.18510/ijsrtm.2019.715
Subject(s) - penalty method , mathematical optimization , quadratic programming , particle swarm optimization , mathematics , constraint (computer aided design) , linear programming , quadratic equation , exponential function , meta heuristic , heuristic , function (biology) , computer science , mathematical analysis , geometry , evolutionary biology , biology
Purpose: The present paper focuses on the Non-Linear Programming Problem (NLPP) with equality constraints. NLPP with constraints could be solved by penalty or barrier methods.
Methodology: We apply the penalty method to the NLPP with equality constraints only. The non-quadratic penalty method is considered for this purpose. We considered a transcendental i.e. exponential function for imposing the penalty due to the constraint violation. The unconstrained NLPP obtained in this way is then processed for further solution. An improved version of evolutionary and famous meta-heuristic Particle Swarm Optimization (PSO) is used for the same. The method is tested with the help of some test problems and mathematical software SCILAB. The solution is compared with the solution of the quadratic penalty method.
Results: The results are also compared with some existing results in the literature.