International Monthly Seminar on Time Scales Analysis #9

This talk investigates the conditional Hyers–Ulam stability of first-order nonlinear logistic models, both continuous and discrete. Identifying bounds on both the relative size of the perturbation and the initial population size is an important issue for nonlinear Hyers–Ulam stability analysis. Utilizing a novel approach, for h-difference equations we derive explicit expressions for the optimal lower bound of the initial value region and the upper bound of the perturbation amplitude, surpassing the precision of previous research. Furthermore, we obtain a sharper Hyers–Ulam stability constant, which quantifies the error between true and approximate solutions, thereby demonstrating enhanced stability. The Hyers–Ulam stability constant is proven to be in terms of the step-size h and the growth rate, but independent of the carrying capacity. Detailed examples are provided illustrating the applicability and sharpness of our results on conditional stability.

Autoregressive (AR) models are fundamental tools in time series analysis, capturing temporal dependencies through lagged observations. While traditional approaches often focus on long-term dynamics, many real-world phenomena—such as high-frequency financial data, climate fluctuations, and energy demand—exhibit behaviors that are best understood at shorter time scales and are often influenced by seasonal effects. In this work, we introduce and study the Short Time Scale Autoregressive (STAR) process, designed to model short-range temporal correlations with particular attention to rapidly evolving structures in the data while explicitly incorporating seasonal components.