Improving Real-parameter Genetic Algorithm Based on Simulated Annealing and It's Engineering Application

With its inheritance,genetic algorithm may spend much computation time in the encoding and decoding process.Also,since genetic algorithm lacks hill-climbing capacity,it easily fall in promatureness and local convergence.A novel adaptive real-parameter simulated annealing algorithm(ARSAGA) that maintains the merits of genetic algorithm(GA)and simulated annealing(SA)is proposed.Adaptive mechanisms are also added to insure the solution quality and to improve the convergence speed.This method to solve the helical spring constrained optimization design problem is applied.The results indicate that the global searching ability and convergence speed of this novel hybrid algorithm is significantly improved,even though small population size is used for a complex and large problem.