Investigation of optimal parameters in multiagent dynamical system

In complex multiagent systems, such as those found in large companies, coordinating multiple agents to efficiently complete numerous tasks poses a significant challenge. This paper aims to study the identification of optimal parameters that accurately describes agent preferences, thereby enabling the shortest path to project completion. By mimicking real-world scenarios, we develop a multiagent system, that emulates the organizational structure of a large company. Our research focuses on investigating the feasibility of finding optimal parameters that align agent preferences with project completion time. We investigated different techniques of updating agents α parameter, which is responsible for agents preferences over available tasks, which are distributed among free agents by Stable Improvements Cycles (SIC). The dynamics of systems with different approaches empirically proved the existence of optimal α parameter, which is dependent on agents competence.