Abstract: An optimal matching algorithm based on matrix computation was proposed to solve the problem of how to get the most pairs of two matched thermal resistances with the special matching roles in large amount of resistances. The algorithm can also be applied to other matching problems.  All feasible matched pairs could be got by a cycle of matrix computation. The influence factor of each feasible combination would be drawn from the pairing rate. Selected priority of each feasible combination is determined by the influence factor. The optimal screening process is based on the selected priority so as to obtain the final matching plan. Monte Carlo method was used to validate the performance of the optimal algorithm.

Key words: optimal matching algorithm, matrix computation, influence factor