Hybrid optimization for constrained mixed discrete nonlinear problems: An application to the design of an environmental conscious transonic aircraft

Stephan Lehner, Purdue University

Abstract

Economic and environmental considerations call for highly sophisticated new aircraft designs. Numerical optimization procedures can aid the designer in searching for the best available design. However, an aircraft optimization problem is a mixed discrete nonlinear problem which consists of continuous and discrete design variables. The latter stand for both design choices and the possible application of new technologies that might enable a significant advantage compared to conventional designs. This paper approaches this problem with a hybrid optimization algorithm, i.e. the use of two different optimization algorithms in a single process. Here, this is a combination of a binary-coded genetic algorithm and a gradient based method as the hybrid optimizer. Applying this hybrid optimization method to simple, multi-modal test problem reveals that a Lamarckian strategy and applying the local search to each individual in each generation is a well suited hybrid approach. For the design of a 150-seat aircraft, the hybrid optimizer identified a slightly forward swept design. It operates at low cost and has a reduced environmental impact compared to current aircraft with a similar mission. The algorithm is most sensitive to the natural laminar flow technology model. The hybrid optimizer is then used to explore designs that must satisfy more stringent environmental constraints.

Degree

M.S.A.A.

Advisors

Crossley, Purdue University.

Subject Area

Aerospace engineering|Environmental engineering

COinS