A new neural network approach to economic emission load dispatch

An artificial neural network method is developed for the solution of economic emission load dispatch (EELD) problems with thermal generation. The proposed method can overcome numerical difficulty caused by conventional neural networks with network parameters, and the states of the dynamic system described by the new neural network converge globally to the optimal solution of the EELD problem whenever its initial points are located inside or outside the feasible region of the problem. The application and validity of the proposed algorithm are demonstrated with a sample system with three generators.