Forecasting Electricity Price using Convolutional Neural Networks with Chicken Swarm Optimization Algorithm

Forecasting electricity prices is critical to the efficient operation and planning of power systems, allowing market actors, governments, and consumers to make educated decisions. This abstract provides an overview of a thorough examination and comparison of power price forecasting strategies. This paper proposes a Deep Learning model using Convolutional Neural Networks (CNN) with the Chicken Swarm optimization Algorithm (CSOA) to anticipate changes in the price of power in Australia’s grid. It evaluates them in relation to other cutting-edge models. The CNN architecture is intended to extract hierarchical features from input data by leveraging its capacity to learn spatial and temporal connections. Backpropagation trains the model, adjusting the network’s weights and biases to decrease predicting error. The CSO algorithm improves the optimization process and prevents the CNN from becoming trapped in local optima. CSO uses a population-based search technique to efficiently explore the solution space inspired by the collective behaviors of chickens. We have selected indicators for evaluation that will show how our model outperforms the state-of-the-art approaches at predicting future electricity costs.