SOLVING ENGINEERING OPTIMIZATION PROBLEMS WITH THE SWARM INTELLIGENCE METHODS
An important stage in problem solving process for aerospace and aerostructures designing is calculating their main characteristics optimization. The results of the four constrained optimization problems related to the design of various technical systems: such as determining the best parameters of we...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | Russian |
| Published: |
Moscow State Technical University of Civil Aviation
2017-05-01
|
| Series: | Научный вестник МГТУ ГА |
| Subjects: | |
| Online Access: | https://avia.mstuca.ru/jour/article/view/1052 |
| Summary: | An important stage in problem solving process for aerospace and aerostructures designing is calculating their main characteristics optimization. The results of the four constrained optimization problems related to the design of various technical systems: such as determining the best parameters of welded beams, pressure vessel, gear, spring are presented. The purpose of each task is to minimize the cost and weight of the construction. The object functions in optimization practical problem are nonlinear functions with a lot of variables and a complex layer surface indentations. That is why using classical approach for extremum seeking is not efficient. Here comes the necessity of using such methods of optimization that allow to find a near optimal solution in acceptable amount of time with the minimum waste of computer power. Such methods include the methods of Swarm Intelligence: spiral dynamics algorithm, stochastic diffusion search, hybrid seeker optimization algorithm. The Swarm Intelligence methods are designed in such a way that a swarm consisting of agents carries out the search for extremum. In search for the point of extremum, the particles exchange information and consider their experience as well as the experience of population leader and the neighbors in some area. To solve the listed problems there has been designed a program complex, which efficiency is illustrated by the solutions of four applied problems. Each of the considered applied optimization problems is solved with all the three chosen methods. The obtained numerical results can be compared with the ones found in a swarm with a particle method. The author gives recommendations on how to choose methods parameters and penalty function value, which consider inequality constraints. |
|---|---|
| ISSN: | 2079-0619 2542-0119 |