Simulation of Fuzzy Intelligent Agents in Stone-Age Virtual Ecosystem
Abstract
Artificial life (A-Life) is a relatively new research paradigm which investigates the life-like systems, their processes, and their evolution, through the use of simulations. Several A-Life simulators have been already presented in the literature. These simulators are mostly designed for studying the behaviors of living organisms based on a world of agents, rules and objects. In these simulations, the agent’s experience along with simple reinforcement learning algorithms are used to produce complex behaviors. However, such worlds are not suitable for cognitive studies focused on complex human-like behaviors, mostly because of their slow calculation speed and discrete model structures. This paper proposes a new simulator, called Stone-age, for simulating and evaluating the behavior of the agents in the environment. Generally speaking, Stone-age simulates a form of ecosystem, or an environment of life, in which a series of primitive human beings are living in an artificial world with prehistoric creatures and objects. Due to its flexible structure and complex learning algorithms, we believe that Stone-age is a suitable tool for cognitive simulations in studying the emergence of complex human-like behaviors. This study is mainly focused on approaches and algorithms that can enhance the knowledge level of the agents and improve their decision-making process.
Keywords
Fuzzy Logic, Artificial Life (A-Life), Evolutionary Reinforcement Learning, Stone-Age Ecosystem, Intelligent Agent, Multi-Agent Environment