Defect and variation tolerance logic mapping for crossbar nanoarchitectures as a multi-objective problem

Abstract

In addition to high defect density, serious variations exist in self-assembled nanoelectronic crossbar architectures compared to the conventional lithography-based CMOS integrated circuits. Therefore, it is one of the emerging challenges to tolerate variations apart from tolerating defects. In addition, considering many-sided factors simultaneously in variation tolerance is necessary. In this paper, we first consider two important factors in variation tolerance simultaneously and formulate this problem as a multi-objective optimization problem, then employ and extend an effective multi-objective evolutionary algorithm called nondominated sorting genetic algorithm II (NSGA-II) to the variation tolerance. Experimental results show the effectiveness of the proposed approach.