Comparison of sGA and SEGA Methods to Solve the Problem of Power Generation and Power Losses on Distributed Generating Systems

This paper introduces a new algorithm to solve the problem of power flow optimization. This algorithm is the Spontaneous Evolutionary Genetic Algorithm (SEGA). SEGA is a combination of Neural Network (NN) and Standard Genetic Algorithm (sGA). SEGA conducts individual selection in a population to get the best results. This individual is a representation of each power station in an electrical network system, because the generator is the object to be optimized. SEGA is different from the previous generation of SGA Standard Genetic Algorithm. SEGA generates new populations (excluding the main population), to get more optimal results. Crossover, mutation and recombination are the selection algorithms used in SEGA, to obtain individual or more optimal results. In this simulation is used IEEE 57 bus plant, as well as compared the quality of the solution between the proposed algorithms with sGA. With this algorithm proven that, SEGA able to improve result (Fitness) from sGA. By using SEGA, the cost of generation on an IEEE 57 bus system is cheaper at 110 $ / hr compared to using sGA.