The results of optimizing some of the geometric design variables of the inletlejector section of a rocket-based combined-cycle (RBCC) engine and the study of its performance trends are presented. A two-dimensional rocket-ejector system was studied over a matrix of engine design variables. Bypass ratio, ejector compression ratio, and ejector/mixer thrust efficiency were used to analyze RBCC internal flowpath physics. The primary thruster exit How properties were calculated with the reacting and multiphase program, which was used as fixed inlet conditions for the ejector/mixer analysis. The computational fluid dynamics simulations of the inlet/ejector system were carried out with a finite difference Navier-Stokes code. The GO(2)/GH(2) combustion physics were solved for finite rate conditions with a system of seven species and nine reactions. Neural networks techniques were used to generate the response surface. The desirability function approach of optimization, tied with the response surface, was used for inlet/ejector optimization. This approach gives the designers the freedom to set up their own priorities on the response values to be built into the optimization procedure. Optimum values of primary thruster size, duct length-to-diameter ratio and ramjet burner to ejector/mixer inlet area ratio were obtained. Engine performance trends as functions of geometric variables were also studied.