TY - EJOU AU - Poongavanam, Elumalaivasan AU - Kasinathan, Padmanathan AU - Kanagasabai, Kulothungan TI - Optimal Energy Forecasting Using Hybrid Recurrent Neural Networks T2 - Intelligent Automation \& Soft Computing PY - 2023 VL - 36 IS - 1 SN - 2326-005X AB - The nation deserves to learn what India’s future energy demand will be in order to plan and implement an energy policy. This energy demand will have to be fulfilled by an adequate mix of existing energy sources, considering the constraints imposed by future economic and social changes in the direction of a more sustainable world. Forecasting energy demand, on the other hand, is a tricky task because it is influenced by numerous micro-variables. As a result, an macro model with only a few factors that may be predicted globally, rather than a detailed analysis for each of these variables, is required. In this work, a hybrid approach is proposed for identifying the optimal generation-based mix of electricity systems in India. This approach is developed by combing Recalling-Enhanced Recurrent Neural Network (RERNN) and Giza Pyramids Construction (GPC). RERNN possesses selective memory features whereas GPC is a meta-heuristic algorithm that deals with different sets of problems. The goal of this approach is to assess the present load requirements and production profile to understand the current requirement and its production methods. Data on regards to electricity load profile involving the generation of all power plants, capacity factors and transmission limits were gathered from different sources including the Indian National Load Dispatch Centre, Central Electricity Authority database, annual reports of India and regional load dispatching centers. The proposal has introduced the RERNN technique to analyze the optimal time series whereas the GPC technique is involved in the optimization of RERNN parameters. The present work is simulated in MATLAB and the performance assessed with different parameters. The present technique gives an effective results in terms of accuracy, computational efficiency and feasibility. KW - Electricity generation; demand; RERNN; GPC; time series; prediction DO - 10.32604/iasc.2023.030101