In the last decade, the integration of Renewable Energy Sources (RESs) in distribution networks has been constantly increasing due to their many technical, economical, and environmental benefits. However, the large-scale penetration of RESs is limited by the grid security constraints, e.g., voltage and current limits. The control of inverter-based RESs can guarantee compliance with those constraints while preserving the RESs performance. However, the installation of additional controllable converter units introduces additional investment costs and has therefore to be limited. In this paper, a Mixed-Integer Second-Order Cone (MISOCP) optimization problem is developed to optimally configure the layout of the Q-V and P-V droop controllers for the RES converters. The controllers' layout is optimized to minimize investment, maintenance, and energy purchase costs subject to the system constraints. To provide an accurate and convex model of these constraints, we adapt an augmented relaxation method, recently proposed to address the optimal power flow problem in radial distribution networks. Our method is tested on a standard IEEE 34-bus network and the results are compared to those provided by standard approaches.