Addressing the 0-1 programming problem focused on multi-UAV (Unmanned Aerial Vehicle) task allocation, an improved Firefly Algorithm (FA) and an improved Genetic Algorithm (GA) are proposed. The enhanced FA adopts a novel position updating strategy, mimicking "superior" individuals while considering mutation factors, and concurrently designs a repair function to ensure constraint satisfaction. The refined GA encodes chromosomes as column vectors and performs repair operations on the iterative population. Simulations comparing the mean, variance, and runtime of the two algorithms under different population sizes reveal that the improved FA can achieve more accurate results when computational resources are sufficient, whereas the improved GA exhibits an advantage when rapid results are required.