Publication

Abstract : Nowadays in electricity markets where purchases and sales are made through bids. Many times, firms may collude, leading to inflation of market prices. A neural network-based approach is introduced for identifying the collusions between different firms in the electricity market. Different collusion scenarios between the firms are identified initially. For every method, market equilibrium and demand for load are calculated. In this work, market equilibrium points are used for generating the dataset for four different firms with ten test generators. The accuracy in finding out the collusions is evaluated, and the accuracy for 1D CNN is compared with CART and SVM models. 1D CNN demonstrates the best accuracy of 99%.