This article presents both the development and the application of a hybrid meta-heuristic algorithm, Particle Swarm Optimization + Simplex (PSO+SX), and seeks to evaluate its performance in the solution of heat transfer problems shown by electronic circuits. More specifically, the minimum entropy generation criterion for the design of micro-channels was used. In order to check the validity of the algorithm, this work used data from recent literature, where deterministic algorithms were implemented. It was possible to reproduce the results with a lowered computational cost (through simpler programming,) which allows the designer to focus on the heat problem instead of the numerical solution. The proposed method is better than Lagrange multipliers since it does not require starting conditions near the global minima to ensure convergence and precision, thus making it a feasible technique for real-life problems.