Energy industries face the challenge on how to design networks to gather flows from unconventional assets due to the relative short lifetime of the wells. The optimal design of the network of pipelines and processing facilities is a challenging problem from both combinatorial and nonlinear viewpoints. To date, real instances of this problem cannot be solved to global optimality unless simplifying assumptions are made. We propose a systematic algorithm to address the optimal planning of gathering networks in a multiperiod horizon. It relies on the strategic selection of links on which fluid-dynamic equations are imposed. By relaxing constraints, the resulting mixed-integer programming models are still complex from the combinatorial standpoint, but they can be solved in reasonable computational times. The algorithm progressively adds constraints to the potential arcs, seeking for the global optimal solution by selectively tightening relaxations. Results show that computational times can be reduced by two orders of magnitude.