2016 Elsevier B.V. MittagLeffler stability analysis of fractional-order neural networks with variable-time impulses is addressed in this paper. Several well-proposed conditions with theoretical demonstration ensuring that every solution of concerned models intersects each surface of the discontinuities exactly once are provided. Meanwhile, by applying B-equivalence method, the reduced fractional-order neural networks with fixed-time impulsive effects can be regarded as the comparison systems of the investigated original network models. Furthermore, a series of sufficient criteria illustrating the same stability properties between both variable-time impulsive fractional-order neural networks and the fixed-time alternative, and guaranteeing the stability of the considered models are presented. Finally, two simulation examples are presented to demonstrate the efficiency and feasibility of the achieved results.
