A general finding solution method is introduced to high dimension multi-objective hybrid discrete variables and the multi-objective is transformed into single object with relative degree of grey incidences

and then the optimal solutions are found by improved particle swarm optimizer （ PSO）. The grey improved PSO algorithmic program for the multi-objective optimization of hybrid discrete variables is developed. The method can reasonably deal with value adopting problems of hybrid discrete variables in optimization design. The PSO algorithmic has been improved and dynamical penalty function is adopted to transform the constrained optimization problems into unconstrained optimization problems. The example of multi ?objective optimization design shows that this algorithm has no special requirements on the characteristics of optimal designing problems

it has a fairly good universal adaptability and a reliable operation of program with a strong ability of overall convergence.