You also need to provide the cost function and I do not agree that using mutation or not depends on the population size. If your aim is just to understand the behaviour of a standard GA when changing the parameters, ok.
If you need to make any serious application of a GA to a complex problem, much effort than that is needed (maybe implying the design of a new solution coding and the proper mutation and crossover operator).