Metaheuristics approaches to solve combinatorial optimization problems in distribution power systems. An application to Phase Balancing in low voltage three-phase networks

• We present an innovative approach to solve multiobjective Phase Balancing problem.
• This problem cannot be solved satisfactorily by classic optimization methods.
• The proposed approach is based in metaheuristics algorithms and fuzzy sets.
• Two metaheuristiscs, FEPSO and FSA, are introduced to compare its results.
• On a real system simulation, the FEPSO metaheuristic has the best performance.

Metaheuristics algorithms are widely recognized as one of most practical approaches for combinatorial optimization problems. One the most interesting areas of application are the power systems. In particular, Distribution Systems planning and operation. This paper presents some metaheuristics approaches to solve a typical combinatorial optimization problem: the Phase Balancing in Low Voltage Electric Distribution Systems. A model supported in Linear Integer-Mixed Programming is presented, to observe and discussing its limitations. From this, is introduced a new metaheuristic, called Fuzzy Evolutionary Particle Swarm Optimization, based in the Swarm Intelligence Principles and Evolution Strategies, which is extended to fuzzy domain to modeling a multi-objective optimization, by mean of a fuzzy fitness function. A simulation on a real system is presented, and advantages of this approach respect to the Classical Simulated Annealing and Particle Swarm metaheuristics, selected between the most representatives, are evidenced.