Badanie skuteczności algorytmu nieliniowego simpleksu z leksykograficznym sposobem oceny rozwiązań w optymalizacji silników indukcyjnych

The paper presents modified nonlinear simplex algorithm with lexicographic order comparison of produced solutions and its application to the optimal induction motors design. In the comparison of solutions produced in optimization process have been taking into account both the objective function value and the additional parameter called satisfaction level of constraints. This method of comparison allows advantage of feasible solutions and control of the degree of the advantage during the optimization process. The special attention was pointed out on the choice of algorithm parameters and kind of mutation operator. Presented algorithm have been implemented in object-oriented software. Calculation results of selected double-cage induction motors have been compared with results obtained by application evolution strategy (μ+λ)-ES and hybrid algorithm assembled with this strategy and modified Price algorithm. The additional calculation experiment allows the comparison of exploitation properties between α-Constrained Simplex Method and Modified Price Algorithm. The examinations demonstrates, that the algorithm can be used in optimization of induction motors, but only by not very restrictive constraints concerning functional parameters.