Using an evolutionary algorithm for catalog design |
| |
Authors: | Susan Carlson-Skalak Michael D White Yong Teng |
| |
Affiliation: | (1) Department of Mechanical, Aerospace, and Nuclear Engineering, University of Virginia, Thornton Hall, McCormick Road, 22903-4221 Charlottesville, VA, USA |
| |
Abstract: | This paper describes an evolutionary algorithm that was developed for catalog design. This algorithm is based on genetic algorithms, but uses an object-oriented coding scheme to represent a design, and introduces unique crossover and mutation operators. To account for the dependence of system performance on both system configuration and component selection, the evolutionary algorithm allows for simultaneous alterations of configurations and components. This new approach allows the consideration of alternate configurations and allows the configurations to evolve to make the best use of the available components. Using this evolutionary algorithm, a piping system was designed in which cooling fluid was delivered to three machines on a manufacturing floor at specified pressures and flow rates. The algorithm was able to find good designs that satisfied the given design specifications. |
| |
Keywords: | Catalog design Configuration design Genetic algorithm |
本文献已被 SpringerLink 等数据库收录! |
|