A Hybrid Multi-Algorithmic Strategy for Real-Time Forecasting and Economic Optimization in Renewable Microgrids

Micro grids powered by renewable energy face challenges in balancing variable generation with fluctuating demand, often resulting in operational inefficiencies and higher energy costs. Accurate forecasting of load, solar generation, and tariff trends is essential to enhance reliability and minimize cost. This paper presents a comprehensive hybrid forecasting and optimization framework that integrates statistical models-ARIMA, SARIMA, and Holt-Winters with a Multi-Algorithmic Strategy (MAS) hybrid approach. The proposed methodology was validated on a real micro grid dataset from April 2025, with the objective of reducing the average purchase cost of electricity while ensuring operational feasibility. The hybrid MAS model demonstrated superior forecasting performance across all variables, achieving lower error measures compared to individual models. In addition, economic analysis showed that the hybrid framework reduced the average unit price of energy procurement from 4.22/kWh (baseline) to 3.71/kWh, corresponding to a 12% cost reduction. This improvement was achieved under practical battery storage and power exchange constraints, highlighting the framework’s adaptability to real-world micro grid settings. By jointly addressing forecasting accuracy and economic optimization, the proposed framework provides a scalable solution for enhancing the sustainability and cost-effectiveness of renewable energy micro grids. The outcomes offer valuable insights for policymakers, utilities, and researchers seeking to advance micro grid planning and operation in the context of increasing renewable penetration.