AN EXPERIMENTAL MULTIOBJECTIVE OPTIMIZATION

The application of multiobjective programming in experimental investigations was studied. The minimax method and the method of sequentially built compromise function were proposed. The investigations were carried out in a sequential way using the simplex method. To illustrate the experimental multiobjective methods, the data concerning the flow of liquid in Kenics Static mixers were used. A compromise was investigated between the pressure drop and degree of mixing. For the proposed weights equal w1 = 0.2 for DELTA-p and w2 = 0.8 for M, respectively, such a compromise was obtained using a the minimax method that the number of mixing elements was 20 at flow velocity 0. 17 m/s. In the case of optimization by the method of sequentially built compromise function, it was assumed that the feasible value of M should be equal to 99% of the value of M(max). Then, the compromise solution yielded N = 16 elements and flow velocity u =0.1 m/s.