Reaching Long-Term Goals Based on Local Interaction between an Organism and the Environment: Computer Simulations Based on Adaptive Behavior

An agent-based computer simulation created using the adaptive-behavior approach has shown that in some cases it is possible to reach a long-term goal (defined here as a desirable place far removed from the agent in time and in space) without necessarily having an exhaustive, predefined plan of action, but rather by using a set of rules on sensorimotor couplings that govern the organism's local interaction with its environment. We used an agent-based simulation program that emulated a two-dimensional microworld in which agents interacted with their environment and other agents through a set of sensorimotor couplings to reach a long-term goal. The main hypothesis was that the sensorimotor coupling, which consisted of following other agents, led to the emergence of collective behavior that enabled the agents to attain the long-term goal. We systematically varied the sensorimotor couplings that allowed agents to follow other agents, agents' perceptual field length, and the number of obstacles. In total, there were 160 independent simulations for each design cell. Analysis supported the hypothesis and suggested that, in addition to the process of reaching long-term goals, other cognitive processes, such as categorization, memory, and reasoning, can be reconsidered bearing in mind the emergence of behavior and an approach based on interaction between the organism and its environment.