Using genetic algorithms as an automated methodology for conceptual design of rotorcraft.

摘要

Genetic Algorithms (GA) are a relatively new class of search and optimization algorithms. Following the "survival of the fittest" patterns of natural selection and reproduction displayed by biological populations, the capabilities of GA as a non-optimization method make them particularly well suited calculus-based global search and to perform conceptual design tasks. This dissertation presents an introduction to genetic algorithms and a discussion of their application to conceptual design. A simple GA has been combined with an industry standard sizing code to perform the conceptual design of helicopters. Using this approach, appropriate variants of the GA's operators and values of control parameters were determined for performing helicopter design tasks. A final form of the genetic algorithm-based design code was then used to generate conceptual designs of the Medium Lift Replacement Helicopter and a notional attack helicopter. Results of this research are discussed, providing insight into both the genetic algorithm construct required to perform conceptual design and the actual ability of a genetic algorithm to perform as an automated conceptual design methodology.