To minimize the comprehensive power generation cost of microgrids

this article undertakes a study focusing on the optimization configuration of microgrid capacity

employing the improved sparrow search algorithm （ISSA）. A capacity optimization model is formulated

with the annual comprehensive cost of microgrids serving as the objective function. This model incorporates various factors

such as price-based demand response

the levelized cost of energy storage systems

and time-of-use electricity pricing. Addressing the challenges of inadequate convergence accuracy and susceptibility to local optima in the traditional sparrow search algorithm

this study introduces an enhanced version of the algorithm. Initially

a chaotic reverse learning strategy is adopted to ensure a uniform distribution of the initial population. Subsequently

the golden sine formula is integrated to enhance the algorithm's convergence accuracy. Additionally

dynamic weight factors are introduced to improve the algorithm's dynamic performance and bolster its global search capabilities. Lastly

a mixed mutation perturbation strategy and a greedy strategy are employed to facilitate escaping from local optima. To validate the proposed ISSA

a comparative test is conducted against the sparrow search algorithm

grey wolf algorithm

and whale algorithm. The results demonstrate that ISSA exhibits superior convergence accuracy and faster convergence speed. Utilizing actual data from a specific region for numerical simulation

the calculated comprehensive power generation cost is reduced by 6.39%

7.35%

and 11.68% when compared to the other three algorithms

respectively. This underscores the effectiveness and superiority of the ISSA in optimizing microgrid capacity configuration.