Large-Scale Network Layout Optimization for Radial Distribution Networks by Parallel Computing

This paper presents an algorithm to simultaneously optimize the layout and conductor type in a radial distribution network. The optimization includes the investment cost and losses in the lines, with the maximum current constraints per conductor, maximum voltage drop in any node in the network, and tapering constraints. A branch-exchange algorithm is used for layout optimization; this generates intermediate points to avoid reaching local minimums, and conductor optimization is solved with a dynamic programming algorithm. There is also a variant for use in large-scale areas with several feeding points and a set of geographically distributed loads, which do not require a preassignation of the loads, allowing their connection to the point that offers lower overall cost. To achieve reasonable resolution times, a grid is built to utilize parallel computing.