| Sign In to gain access to subscriptions and/or personal tools. |
A Novel Genetic Algorithm Based on Gene Therapy TheorySchool of Computer Science and Engineering, Xian University of Technology, Xian 710048, China, Wbllw{at}126.com
School of Computer Science and Engineering, Xian University of Technology, Xian 710048, China
College of Computer & Information Engineering, Hohai University, Nanjing 210098, China
School of Computer Science and Engineering, Xian University of Technology, Xian 710048, China Based on the analyses of a genetic algorithms properties and shortages, a novel genetic algorithm (GTGA) is proposed with analogies to the concept and method of gene therapy theory. The core of GTGA lies on construction of a gene pool and a therapy operator. The gene pool, which is first created according to prior knowledge and then is updated according to posterior knowledge, contains eminent genes, morbid genes and their character istic information as well. The therapy operator consists of an insertion operation that inserts eminent genes into an individual and a removing operation that removes morbid genes from an individual. The methods of creation and updating of the gene pool and contruction of the therapy operator are given and demonstrated by the travelling salesman problem (TSP). To validate the superiority of the GTGA, a conventional genetic algorithm (GA), a novel genetic algorithm based on immunity (IGA) and the GTGA are compared as regards TSP. The simulation results show that the GTGA can restrain the premature convergence phenomenon effectively during the evolutionary process while greatly increasing the convergence speed.
Key Words: genetic algorithm gene pool gene therapy therapy operator traveling salesman problem
Transactions of the Institute of Measurement and Control, Vol. 28, No. 3,
253-262 (2006) |
|||