Multiobjective Optimization of Two-Stage Rockets for Earth-to-Orbit Launch

Rockets are defined by many variables and constraints, and ultimately deliver a payload to orbit at some cost. These characteristics provide the basis for an optimization problem. This paper describes the construction and optimization of a two-stage liquid rocket model. At the core of the model, the shooting method was used to calculate the fuel usage and ending altitude of the rocket design. The design space was examined using a multi-objective genetic algorithm. A wide range of feasible non-dominated rocket designs were found and these are presented in the paper. Sensitivity to various input parameters was also explored. The model will serve as a benchmark problem for future research and comparative analysis of a number of optimization algorithms.