Font Size: a A A

The Study Of The Tabu Search Algorithms And Its Application To The Electromagnetic Field Optimizations

Posted on:2004-09-08Degree:MasterType:Thesis
Country:ChinaCandidate:X H ZhangFull Text:PDF
GTID:2132360092986277Subject:Electrical theory and new technology
Abstract/Summary:PDF Full Text Request
The complexities of modern engineering optimization problems are their nonlinearity and multi-peak behaviors. They are so complicated that in some cases it is hard/impossible to describe their mathematical functions. That they are easy to fall into a local optimum is the shortcoming of conventional optimization methods. Modern algorithms are supposed to be able to solve this problem effectively. However, compared with conventional methods, they are too time-consuming. Therefore, researchers have been interested in improving them so as to raising the searching speed. This thesis focuses on Tabu search algorithms used for the function optimizations. The improvement ideas are proposed and the effectiveness is shown with the testing of both theoretical mathematical functions and a practical magnetic filed optimization problem.The idea of the improvements is to take advantage of genetic algorithms. Below are the details:1. Perform mutation operations on the promising areas. These operations make the searching move to more areas, which might have the optimum solution inside.2. Perform crossover operations on the present solutions. These operations form new solutions, which have the information from more than one available solution. This makes it possible to find the solution faster.3. Improve the management method of the promising list. There are two lists used in the improved intensive searching process. One is for storing the promising areas; the other is for their combination, selection and adjustment. And the length of the promising list can be adjusted automatically according to the number of the promising areas. In this way, the computation cost can be reduced.4. Improve the ending criteria. The ending of the improved searching depends on the number of the terms left in the proposing list and the value of the objective functions.The above improvement procedures were tested with functions of several variables and multi optimum solutions and they performed well.For validation in engineering application, the proposed method was applied to Team Workshop problem 22 and the results obtained were satisfactory.
Keywords/Search Tags:Tabu Search Algorithms, Enhanced Continuous Tabu Search Algorithms, Genetic Algorithms, Superconducting Magnetic Energy Storage
PDF Full Text Request
Related items