For the 0-1 programming problem of multi-UAV (Unmanned Aerial Vehicle) task allocation, an improved Firefly Algorithm (IFA) and an improved Genetic Algorithm (IGA) are proposed. The improved Firefly Algorithm adopts a new position update strategy, mimicking "excellent" individuals while considering mutation factors, and a repair function is designed to ensure constraint satisfaction. The improved Genetic Algorithm encodes chromosomes as column vectors, performs repair operations on the iterative population, and compares the mean, variance, and runtime of the two algorithms under different population sizes through simulations. Simulation results indicate that the improved Firefly Algorithm can obtain more accurate results when computational resources are sufficient, while the improved Genetic Algorithm has an advantage when quick results are needed.