Integrating hysteresis into system models

Abstract

Hysteresis is commonly present in most physical, biological, economic and social systems. In this paper, we show hysteresis in a lattice network of individually non-hysteretic agents by allowing for system memory. We show that hysteresis depends on: (1) agent response to the external input variable, (2) agent response time and (3) interactions among agents. Our simple system exhibits rich behavior including power law variations, rapid fluctuations and system saturation. We then describe the hysteresis and fluctuations of the Greenland Ice Core Project oxygen isotope variation using our model.