Four approaches for groundwater development planning

Groundwater system development problems usually involve the placement and the operation of a set of wells. Finding the optimal solution for such problems, when all the corresponding costs are considered is a very difficult task since non-linear models that include discrete variables have to be solved. This is why such problems are often simplified either by ignoring fixed costs or by linearizing the cost functions. In this paper four different approaches are compared: a non-linear programming with polynomial penalty coefficient method; a non-linear programming with exponential penalty coefficient method; a mixed-integer–non-linear programming commercial solver, and a simulated annealing algorithm, in an attempt to solve this kind of problem while keeping a realistic model formulation not requiring simplifications.