Study On Several Stochastic Global Optimization Algorithms And Application

Many problems in practically all fields of natural science as well technical design, financial planning and travel scheduling involve global optimization problems. The efficient global optimization methods affect the development of these subjects. With the development of the stochastic global optimization methods such as evolutionary computation, simulated annealing and tabu search during 80 ages in the 20th century, some authors study the theory and application of those algorithms and put forward novel algorithms and solve a lot of practical problems. Based on the former research, the author studies the evolution strategies and simulated annealing in the aspects of theory and application. The paper is organized as follows:Part one introduces in detail the status of evolutionary computation, simulated annealing and tabu search in the aspects of theory and application, followed by the main research of this paper.In Part two, it is proved that, the evolution strategy of uniform distribution for continuous functions, asymptotically converges to the global optimal solutions in probability, by using central limit theorem, under suitable conditions. Then a new kind of evolution strategies is proposed. And its convergence property, asymptotically converges in probability, is proved. Numerical results on some typical optimization problems show that the algorithm proposed is efficient. In addition, this new kind of algorithm is applied to medical image registration with good experimental results.In Part three, based on a kind of new simulated annealing, it is proved that, in theory, such algorithms asymptotically converge to the global minimum point with probability one under suitable conditions.The last part concludes the research in this paper and presents the future research on stochastic global optimization algorithms.