Author(s): Shilpa S Babalad | Anand M Shivapuji
Journal: International Journal of Research In Computer Science
ISSN 2249-8257
Volume: 2;
Issue: 2;
Start page: 1;
Date: 2012;
VIEW PDF
DOWNLOAD PDF 
Original page
Keywords: Evolution | Genetic algorithms | parallel programming | system of equations
ABSTRACT
Analysis and understanding of physical systems require modeling the system as set of simultaneous linear/non linear equations and generating solutions to satisfy the system of equations. Analytical approach towards solving the system of equations remains practical so long as considerable constraints are imposed on the modeled system to bring in significant simplicity so as to retain the system model within the scope of defined algorithms for solving system of equations. The current work adopts the concept of genetic algorithm towards evolving a solution to a system of equations. The fundamental strength of genetic algorithms lies in the fact the solution generation is practically unconstrained permitting the methodology to become the superset for all possible realizable problems. One of the oft repeated and highlighted / drawbacks of genetic algorithm i.e. requirements of huge initial population and corresponding extended number of computations and hence time is addressed by exploiting multi-processing capabilities of the current generation hardware as well as system software. Adopted strategy for selecting the best population, implementation flow chart along with a case study is presented.
Journal: International Journal of Research In Computer Science
ISSN 2249-8257
Volume: 2;
Issue: 2;
Start page: 1;
Date: 2012;
VIEW PDF
DOWNLOAD PDF Original pageKeywords: Evolution | Genetic algorithms | parallel programming | system of equations
ABSTRACT
Analysis and understanding of physical systems require modeling the system as set of simultaneous linear/non linear equations and generating solutions to satisfy the system of equations. Analytical approach towards solving the system of equations remains practical so long as considerable constraints are imposed on the modeled system to bring in significant simplicity so as to retain the system model within the scope of defined algorithms for solving system of equations. The current work adopts the concept of genetic algorithm towards evolving a solution to a system of equations. The fundamental strength of genetic algorithms lies in the fact the solution generation is practically unconstrained permitting the methodology to become the superset for all possible realizable problems. One of the oft repeated and highlighted / drawbacks of genetic algorithm i.e. requirements of huge initial population and corresponding extended number of computations and hence time is addressed by exploiting multi-processing capabilities of the current generation hardware as well as system software. Adopted strategy for selecting the best population, implementation flow chart along with a case study is presented.
