Evolutionary Software Agents for Managing Operations in Chaotic Environments

The effective management of dynamic systems that interact with unstable immersions is crucial for a wide range of applications. This includes tasks such as the stabilization of turbulent gas and hydrodynamic flows, chemical and technological processes, signal generation in radio engineering, and asset management in capital markets. The problem lies in the low level of predictability associated with deterministic chaos models, resulting in additional uncertainty derived from random components of observations, as described by traditional statistical models. Numerical methods prove to be the sole viable approach to analyzing the effectiveness of management under these challenging circumstances. In particular, this study focuses on empirical algorithms used for identifying and predicting local trends to develop extrapolation prediction techniques. As the primary example of managing unstable environments, this paper examines the problem of speculative trading in currency markets, which are characterized by stochastic chaos. Unlike physical and technical issues, the currency market is purely informational, and as a result, lacks inertia. Consequently, traditional predicting algorithms implemented by software robots based on reactive control strategies have proven to be ineffective. This study aims to solve this efficiency problem by exploring control strategies based on the optimization of evolutionary parameters in a sequential manner and by approximating the model structure of observation series.