Low cost genetic algorithm to photovoltaic-diesel power system design problem

In order to find ideal design of hybrid photovoltaic-diesel power system, genetic algorithm is an efficacious technique. The optimum design gives architecture structure, which has finest selection of components and size in accordance with suitable controlled strategy to offer budget friendly and dependable energy substitute. During the search of the best solution, among potential solutions, the aforementioned algorithm tries to find out solution that has minimum total net present cost. Yet, while using complex economic models in order to calculate the value of population fitness, this hunt takes the shape of expensive optimization problem. In order to shrink the deficiencies carried by genetic algorithm, current paper proposes low cost version of cluster-based genetic algorithm grounded upon statistical approaches, which considerably reduces the cost for evaluation of fitness function and bolster the performance. Population is divided into several clusters and multiple linear regression model is obtained from each clusters. Principal component analysis takes responsibility to increase possibility of having good estimation. During the course of probing, the values of population fitness are computed from corresponding model that is comparatively cost effective than direct evaluation. Algorithm is being used to identify clusters is denoted by k-means, which operates the process at low cost. The performance of proposed method is judged on the basis of benchmark case study. The obtained results indicate the efficacy of proposed method for the model of hybrid photovoltaic-diesel power design via genetic algorithm workflow