Font Size: a A A

The Improved Ant Colony Algorithm Based On The Optimal Design Of The Engineering Structure

Posted on:2010-12-13Degree:MasterType:Thesis
Country:ChinaCandidate:S J ChenFull Text:PDF
GTID:2178360308990764Subject:Structural engineering
Abstract/Summary:PDF Full Text Request
Structure design and construction are the most important contents in the field of civil engineering,structure construction are based on structure design,thus structural optimization design is the one of the branches researches of importance in structural engineering.As the quickly development of the national economy,engineering structure gets far more complex,the cost goes higher,optimization of engineering structure attracts more and more concerns.However, to other optimization problem,there is no common way to solve.But simplex method can solve the linear programming problem perfectly.Recently,the fast development of computer science bring the boom of the modern heuristic method.Ant colony algorithm was rooted in 90th last century. It got much attention when it was applied in the TSP.But there may go wrong in the ant colony algorithm, such as too much stagnation and time cost. Additively the mold is based on discrete filter. thus there is not much experience in the search of continuous domain.Consequently, text will focus on the following points:(1) To the choose of many parame ters, as the ant colony algorithm built,which was aiming at tsp,it can't reflect the real need of civil engineering.Consequently,the test will discuss respective value range of some important parameters.Aiming at the bugs of the ant colony algorithm,through the improvement of some more ways to gain the ability of more local search and global search,ant colony algorithm of local search and ant colony algorithm of full-stress appeared as a consequence. Ant colony algorithm of local search was much better after comparison.Because the traditional ant colony algorithm is based on discrete filter. There is a real need to improve the application of continuous domain.In nature,pheromone have feature of continuous distribution,the shorter the distance of center,the bigger pheromone is and vice versa.So pheromone has the property of normal distribution.According to the model we built,with the build of many optimizing system of ant colony algorithm and some thoughts of genetic algorithm and tabu search,there was much application to the question of continuous domain. (2) The mathematical model of continuous domain and discrete filter apply ten truss and twenty-five truss.Through the comparison of it and other algorithm, performance of the mathematical model can be proved.(3) With numerical simulation to truss though the use of ansys and the analysis of displacementand stress,we can observe the deformation and dangerous area from the picture.
Keywords/Search Tags:optimization, ant colony algorithm, numerical simulation, truss
PDF Full Text Request
Related items