A semi-inverse design optimization method for the film-cooling arrangement of high-pressure turbine first-stage vanes is initiated based on a combinatorial optimization algorithm, a one-dimensional heat conduction model, and computational fluid dynamics methods, in which inlet temperature distortion, radiation, and inlet swirl are all considered simultaneously. This semi-inverse design optimization method can optimize the total coolant amount of the film-cooling structure while ensuring an acceptable metal temperature distribution, which finally provides a scattered and nonuniform arrangement of the film-cooling holes and a minimal coolant amount. The optimization methodology is tested on the General Electric energy-efficient engine first-stage vane under a high thermal load, and the optimization result is verified by the conjugate heat transfer computational fluid dynamics simulations. As for the optimized cooling structure, a significant improvement of cooling performance is observed while the total coolant amount is slightly reduced compared with the prototype. It is also found that neglecting each of the three factors (inlet temperature distortion, radiation, and inlet swirl) could result in a significantly different film-cooling arrangement while maintaining the overall cooling performance, which highlights the capability of the semi-inverse design optimization method at various design conditions. Nomenclature A = area of outer surface BR = blowing ratio c p = specific heat at constant pressure DR = density ratio d = thickness h = convective heat transfer coefficient k = turbulence kinetic energy/number of sample points on pressure side and suction side M = total coolant mass flow m = mass flow N = population size (of genetic algorithm) n = number of candidate holes p = pressure p = total pressure q = wall heat flux T = temperature T = total temperature Tu = turbulence intensity x = bit of the design variable α = swirl angle η = film-cooling effectiveness λ = heat conductivity ω = turbulent specific dissipation rate/rotation speed Subscripts ave = area-averaged C = child (of genetic algorithm) c = coolant/with cooling/coolant side c1 = coolant inlet 1 c2 = coolant inlet 2 f = film cooling g = hot gas in cascade passage/gas side i = vane cascade inlet max = maximum value o = rotor cascade outlet P = parent (of genetic algorithm) rad = radiative w = wall (metal beneath thermal barrier coating layer)