In this study, the laminar unsteady MHD (magneto hydro dynamic) flow over a stretching rotating disk is investigated. A hybrid algorithm including artificial neural network (ANN) and artificial bee colony (ABC) for optimization of entropy generation in this flow is applied. An ANN model is developed for predicting entropy generation. Approximately, 97 data sets obtained from the published literature, are applied to develop the ANN model. In addition, the entropy generation is set as the fitness function for the ABC algorithm. Finally, results are compared with these of evolutionary optimization algorithms, Particle Swarm Optimization (PSO). All the experiments of this paper have demonstrated that the optimal entropy is 8.41002 kJ/kg at Unsteadiness Parameter(S) =0, Magnetic Interaction Parameter (M) =0 and Reynolds number (Re) =5.
