Multi-objective optimization using a hybrid approach for constrained Mixed Discrete Non-Linear programming problems—applied to the search for greener aircraft

Satadru Roy, Purdue University

Abstract

Recent trends indicate growing interest in the field of Multi Disciplinary Design and Optimization to address complex Mixed Discrete Non-Linear engineering design problems. The work here presents a hybrid multi-objective algorithm and demonstrates its ability to find solutions for a constrained multi-objective mixed discrete non-linear programming problem. The hybrid algorithm uses a Genetic Algorithm as a global search tool with a gradient based Sequential Quadratic Programming algorithm for local search in a way that seems to overcome the demerits of these two algorithms when used independently. The approach here addresses some of the issues that current state-of-the-art optimization techniques face. Handling constraints is a primary concern for most of the optimization algorithms that seek to address mixed discrete non-linear programming problem. Hybridizing two algorithms has proven to outperform their individual counterparts. However, not much is exploited from the process of hybridizing two algorithms other than the computational efficiency of the gradient-based algorithm and exploring capability of the global search algorithms. The work here presents a compatible hybridization between GA and SQP with improved information sharing between the two algorithms. The hybrid approach is later implemented and used to solve a greener aircraft design problem. Pursuing "greener aircraft" with lower emissions and noise than today's commercial transport aircraft has become an important effort across government, industry and academia. A commonly held perspective of pursuing greener aircraft is that a broad suite of new technologies and, potentially, new aircraft configurations provide the means to attaining a greener aircraft rather than incremental improvements to existing designs. Determining the appropriate combination, or portfolio, of technologies requires a method that can both sort through the myriad possible combinations of available technologies and aircraft configurations along with determining the size and dimensions of the best aircraft for a given selection of technologies. Characterizing an aircraft as "greener" requires consideration of several different metrics (e.g. carbon emissions - as measured by fuel burn, nitrogen oxide (NOx) emissions) along with basic economic considerations (e.g. required yield or ticket price). This combination of features makes this a multi-objective aircraft design optimization problem with both discrete and continuous design variables. Applied to the greener aircraft problem, the hybrid multi-objective algorithm seeks to arrive at the best trade-offs between representative environmental and economic metrics. The work here also describes the development of the aircraft sizing tool that uses an integrated analysis framework to evaluate the objective function and the constraint values. While the detail and fidelity of the aircraft sizing model limits the quality of the results, the application suggests that the hybrid algorithm does have promise to assist decision-makers in choosing the appropriate technology portfolio.

Degree

M.S.A.A.

Advisors

Crossley, Purdue University.

Subject Area

Engineering|Aerospace engineering

COinS