I think here g as the optimization target shouldn’t be seen as a variable and g = \sum_i t_i \mathrm{fee} as the goal function shouldn’t be seen as an equation. So it’s better to say: the m + 1 basic variables can be computed from the free variables using the m + 1 equations