Parent Genetic Algorithm Theory And Applications

After analyzing and studying the deficiency of traditional genetic algorithms in solving ordinal combination optimal problem and complex optimal problem with constrains, an improved genetic algorithm is proposed, that is Partheno-Genetic Algorithm (PGA).The Partheno-Genetic Algorithm that repeals the crossover operators of traditional genetic algorithms while uses the reproduction manner with only parent. Comparing with traditional genetic algorithms, the genetic operation of partheno-genetic algorithm is simpler; it is more easy to handle constrains during genetic operation; the initial population of partheno-genetic algorithm need not be varied; there is no "immature convergence" of partheno-genetic algorithm; the searching efficiency of partheno-genetic algorithm is higher.The partheno-genetic algorithm mainly uses serial number strings as solving combination optimal problems, or uses real strings as solving complex optimal problems with constraints.The genetic operators of partheno-genetic algorithm using serial number strings include two classes: gene recombination operators and gene leap operators. The gene recombination operators include gene exchange operators, gene shift operators, gene inverse operators and so on.The function of gene recombination operators of partheno-genetic algorithm is same as the one of crossover operators of traditional genetic algorithms. The evolution mechanism of partheno-genetic algorithm is same-as the one of traditional genetic algorithms. The partheno-genetic algorithm with optimal maintaining operation is global convergent. The partheno-genetic algorithm has hidden parallel as traditional genetic algorithms.Applications of partheno-genetic algorithm using serial number strings to traveling salesman problem, flow-shop and pattern clustering are researched. The simulation examples show the partheno-genetic algorithm is very efficient.For unit commitment of power system, a partheno-genetic algorithm using real strings is given, and its efficiency is proved by simulation examples.