Behavior analysis of new bio-inspired metaheuristics to solve distribution network reconfiguration problem under different radiality constraints representation

摘要

As distribution networks operate in their majority in a radial way, this condition becomes one of the most important constraints of distribution network reconfiguration (DNR) problem. There are diverse ways to represent this in DNR, but a lack of studies that provide an analysis of its behavior is noted, specially regarding its influence in metaheuristics and bio-inspired metaheuristics. This paper presents a discussion between two alternatives to represent this in DNR in conjunction with three new bio-inspired metaheuristics. One is based on the formation of the incidence matrix of the system and its determinant; the other is a modified approach of a method that establishes a set of forbidden switches through the analysis of the system fundamental loops. To solve DNR, three new bio-inspired metaheuristics are presented: monarch butterfly optimization (MBO), gray wolf optimizer (GWO), and marine predators algorithm (MPA). A comparison is performed with selective particle swarm optimization (SPSO). To test the alternatives, three systems are used (33-, 69-, and 84- bus), with results showing the influence of the radiality approach in the results of bio-inspired metaheuristics.