High heat flux is a critical factor limiting the performance and reliability of miniaturized high-power microelectronic systems. In this study, a liquid metal (LM)-based microfluidic cooling system is proposed, with a data-driven computational framework incorporating a learning-based genetic algorithm (LC-GA), aiming to provide efficient thermal management for high-density integrated systems. By integrating LM near-junction cooling with microchannel heat dissipation on a silicon substrate, a heterogeneous three-dimensional interconnect cooling architecture is developed, enabling thermal performance optimization through algorithm-guided parameter tuning. To validate the approach, four distinct microchannel configurations were designed, fabricated, and experimentally tested. After introducing LM into the channels, experimental cooling tests and thermal performance simulations were conducted on a simulated heat source. The experimental results demonstrate that the optimized LM microfluidic cooling system, with parameters determined computationally, can effectively dissipate heat from chips with power levels up to 800 W while maintaining stable thermal performance. Furthermore, multi-factor sensitivity analysis and multi-objective optimization combining response surface methodology (RSM) with the improved LC-GA were performed, enabling the automatic determination of optimal design and operating parameters that balance thermal resistance and pressure drop. The optimized configuration reduced the peak chip temperature to approximately 357.54 K, lowered the system pressure requirement, and increased the performance evaluation criterion (PEC) to 2.327. This study presents a data-driven optimization scheme that supports the development of high-performance integrated microsystems through algorithm-aided thermal design. The results were published in Engineering Applications of Artificial Intelligence under the title “Liquid metal microfluidic cooling system for high-efficiency thermal management via learning-based genetic algorithm”.
