LAUSR

Abstract In this paper, the global dissipativity problem is investigated for the quaternion-valued memristor-based neural networks (QVMNNs) with proportional delay. Firstly, the mathematical model of QVMNNs is proposed, which is a challenging model due to the unbounded delay and the quaternion field. Secondly, based on matrix measure method and Lyapunov theorem, combing with modulus inequality technique and generalized Halanay inequality technique, some sufficient conditions guaranteeing global dissipativity of QVMNNs with proportional delay are derived. Moreover, the global attractive set of QVMNNs can also be obtained. Finally, three numerical examples are presented to demonstrate the effectiveness of the theoretical results.