Evaluating Perception Filters in BDI Jason Agents

Simulation systems are of great help due to their ability to reproduce the real world. In some open scenarios, designers must build intelligent autonomous agents who evolve within and interact with a simulated environment, constrained by a time limit, i.e., In every cycle, agents may deliver an action within a predetermined response time limit. On one hand, the use of Jason interpreter to create and implement intelligent agents can greatly reduce the effort required for the development of BDI agents. However, as agents become more sophisticated, the processing time required for the agents to perceive their environment and to reason about their actions also increases, and in some cases this can exceed this time limit. In this work, we use some performance analysis techniques to measure the effects of including perception filters in the Jason reasoning cycle, aiming to verify whether it causes a significant decrease in reasoning time and if this reduction diminishes the utility of the agents´ actions.