To solve the problems of Multi-objective optimization(MOO) problems in the engineering calculation, the paper proposes a new MOO algorithm named self-adaptive mixed-variable multi-objective ant colony optimization algorithm(SAMOACO(mv)) which uses a self-adaptive parameters deployment and can solve mixed-variable MOO problem. The proposed algorithm begins with designing two indices named non-inferior step and crowded degree to evaluate of the quality of solutions, then, improves ACO(MV) algorithm which can only solve mixed-variable single objective optimization problems by ranking solutions in the solutions archive according to their quality; meanwhile, increases parameters list archive and designs its updating rules referring to the solutions archive; furthermore, the paper designs two mixed-variable MOO benchmark problems based on Kursawe problem and two bar truss problem for purpose to test and compare the functions of SAMOACO(MV) algorithm. The SAMOACO(MV) algorithm is implemented and experimented. Empirical results indicate SAMOACO(Mv) algorithm can find bigger range and more uniform solutions, and can avoid fall into local optimum. The practice shows that the synthetic function of SAMOACO(MV) algorithm is outstanding, and the algorithm is suitable to solve mixed-variable multi-objective optimization problems.
