You are tasked with implementing the main loop of a genetic algorithm to find the maximum value of a mathematical function. The function to optimize is a simple parabola: f(x) = -(x - 3)² + 10. This function has a clear peak at x = 3, where its value is 10.

All the helper functions (init_population, fitness_function, tournament_selection, arithmetic_crossover, mutate) and the main loop structure are provided for you.

Your task is to fill in the core logic of the evolutionary process:

1. Inside the main for loop, you must first evaluate the entire population by applying the fitness_function to each individual. Store these scores in the fitness list.
2. Find the index of the best-performing individual in the current generation and store it in gen_best_idx. (Hint: np.argmax() is useful here).
3. Inside the while len(new_population) < POP_SIZE: loop, you must create a new individual by:
   • Selecting parent1 using the tournament_selection function.
   • Selecting parent2 using the tournament_selection function.
   • Creating a child by combining the parents with the arithmetic_crossover function.
   • Applying variation to the child using the mutate function.
4. After the while loop, replace the old population with the new_population to complete the generation.