Application of evolutionary computation for machine design optimization

Machine optimizations are characterized by many non-commensurable and often competing objectives to be optimized. This paper presents an evolutionary algorithm (EA) based software environment for multi-objective optimal desig n of electromagnetic devices. The main consideration is to produce a strong , inexpensive and lightweight structure that is capable of giving the requi red performance. The software environment allows the users to choose from a variety of EA parameters. The paper presents the results obtained from var iation of several combinations of these parameters tested on a salient mach ine design problem.