Evolutionary Multiobjective Design of Combinational Logic Circuits.pdfVIP

Evolutionary Multiobjective Design of Combinational Logic Circuits Carlos A. Coello Coello Arturo Hern´andez Aguirre Laboratorio Nacional de Inform´atica Avanzada EECS Department R´ebsamen 80, A.P. 696 Tulane University Xalapa, Veracruz, M´exico 91090 New Orleans, LA 70118, USA ccoello@xalapa.lania.mx hernanda@eecs.tulane.edu Bill P. Buckles EECS Department Tulane University New Orleans, LA 70118, USA buckles@eecs.tulane.edu Abstract found this n -cardinality representation scheme more robust than the traditional binary representation used with GAs) In this paper, we propose an evolutionary multiobjective [3, 4]. Our approach presents great (coincidental) resem- optimization approach to design combinational logic cir- blance with the one proposed by Miller [17] and further de- cuits. The idea is to use a population-based technique that veloped by Miller and his colleagues [16, 13]. The 2 main considers outputs of a circuit as equality constraints that differences between the two approaches are the encoding we aim to satisfy. A small sub-population is assigned to scheme and the fitness function. Regarding the encoding, each objective