Application of a genetic algorithm to the optimization of hybrid rockets

A genetic algorithm optimization technique has been successfully applied to the design of a large hybrid rocket booster. Optimizations to minimize gro ss liftoff weight or total inert weight have been carried out using a hybri d rocket sizing code developed at Purdue University. The genetic algorithm was able to find optimal or near-optimal designs that contained both contin uous and discrete variables. Discrete variables included the propellant com bination and the number of fuel ports, whereas the continuous variables, ta nk pressure, chamber pressure, and oxidizer massflux level, were simultaneo usly optimized using the genetic algorithm. Design solutions have been obta ined from a discontinuous design space that contains a very broad, shallow minimum in weight. The resulting designs are discussed with some detail, il lustrating their feasibility and some significant differences with previous ly published designs. The use of a genetic algorithm with ii rocket sizing code appears to offer great potential to designers of rocket systems.