Background: : By developing high-tech instruments, especially digital devices, a need to have efficient and compact cooling is of crucial importance. In the current study, heat transfer enhancement in a microchannel for cooling two hot sources is numerically investigated. Methods: : The effects of imposing a magnetic field, a pulsating flow, and using a hybrid nanofluid are tested. Further, the entropy generation is evaluated and the strength of irreversibility sources is discussed. Reynolds number, Hartmann number, nanofluid volume fraction, inlet velocity amplitude, and frequency are chosen as the independent parameters. An artificial neural network is applied to have a prediction tool for the mean Nusselt number and Bejan number. The analysis for choosing the best neuron number for MLP is conducted. The pre dictive mathematical description is fed to a multi-objective genetic algorithm (NSGA-II) to find the optimum operating conditions of thermal cooling. Significant findings: : The results show that Intensifying the inlet amplitude from 0.1 to 0.8 makes the heat transfer and pressure drop increment respectively by 100% and 300%. However, the amplitude increment remains an inverse effect on the Bejan number by a decrement of about 3–4%. The sensitivity analysis shows the first rank of importance belongs to the Reynolds number the Hartmann number shows the least importance in the ranking.
