The Research Of Filled Function Algorithm For Optimization Problems

For solving global optimization problems, there are many different approaches.Recently, Optimization theory and approach has been greatly developed, because thedevelopment of optimization theory and approach is the requirements in the area ofproduction and living. To find the effective methods for finding the global optimalsolution of general multi-minimizes functions is one of the practical methods for theproblem. The approach of the filled function is an important method of solving globaloptimization problems. The key of this method is to construct a filled function withgood properties. The difficulty of this method is that parameters are too hard to adjust inpractical computation. So, further research is worthy of continuing on exploring thegood filling function. We can construct filled functions with simpler forms, betternatures and more efficient algorithms. This article mainly concerns the course ofdevelopment and the research status of filled function methods for the globaloptimization problems.This paper is divided into five chapters.In the first chapter, an overview of the global optimization model and some of itsmain methods are presented, also the ideas and the development process of the filledfunction are introduced.In the second chapter, according to the ideological and theoretical basis of thefilled function, a filled function with one parameter for solving global optimizationproblem is given an. The excellent nature of this function is discussed. Theoreticalnatures of this filled function are investigated and corresponding algorithm wasdesigned. This function, which contains a single parameter, is easy to adjust in theactual calculation. Finally, the experimental results show that the method is effective.In the third chapter, a new filled function without parameter is proposed for solvinggeneral constraint problem. Then the nature of the function is argued in the paper, a newfilled function algorithm is designed with its theoretical nature. The filled functionwithout parameter is avoid to adjust the parameter.Finally, through the numericalexperiments, this algorithm is proven to be feasible and efficient.In the fourth chapter, this chapter presents a method of auxiliary function that isF-C function. It can solve quickly general unconstrained optimization problems.Compared to the filled function and cross function, there are some difference and the same points in common. The F-C function minimization that needs only one stage oflocal minimization algorithm can get the objective function local minimum that is lowerthan the current minimum. This is a great advantage for F-C function. This paper createsa new F-C function to solve global optimization problems. The properties of the F-Cfunction are discussed and the corresponding algorithm is designed in this paper. Finally,through numerical experiments, we can conclude that the F-C function method isfeasible.The last chapter presents the future prospect of the development of the filledfunction method.