In order to solve the low-efficiency problem of the turbine unit in nuclear power plant(NPP),an evaluation model of a turbine extraction system was established and the corresponding computer program was developed.Based on the computer program,a case study of a 660 MW NPP was implemented to investigate the influence of the extraction parameters on turbine output power.Furthermore,a genetic algorithm and a simplex algorithm were combined to form a hybrid genetic algorithm(GSA).The performance of the GSA was v...