Forecasting and control in overlapping generations model: chaos stabilization via artificial intelligence

TL;DR

This paper demonstrates how artificial intelligence, specifically evolutionary algorithms, can effectively control and stabilize chaotic behavior in an overlapping generations economic model using Pyragas control.

Contribution

It introduces a novel approach combining evolutionary algorithms with Pyragas control to stabilize irregular dynamics in economic models.

Findings

Control successfully stabilizes periodic orbits

Chaotic behavior can be suppressed with optimized parameters

AI techniques enhance analysis and control of complex dynamics

Abstract

Irregular, especially chaotic, behavior is often undesirable for economic processes because it presents challenges for predicting their dynamics. In this situation, control of such a process by its mathematical model can be used to suppress chaotic behavior and to transit the system from irregular to regular dynamics. In this paper, we have constructed an overlapping generations model with a control function. By applying evolutionary algorithms we showed that in the absence of control, both regular and irregular behavior (periodic and chaotic) could be observed in this model. We then used the synthesis of control by the Pyragas control method with two control parameters to solve the problem of controlling the irregular behavior of the model. We solved a number of optimization problems applying evolutionary algorithms to select control parameters in order to ensure stability of periodic orbits. We compared qualitative and quantitative characteristics of the model's dynamics before and after applying control and verified the results obtained using simulation. We thus demonstrated that artificial intelligence technologies (in particular, evolutionary algorithms) combined with the Pyragas control method are well suited for in-depth analysis and stabilization of irregular dynamics in the model considered in this paper.