Proposal for a framework of adaptive mobile intelligent agents

Intelligent agents are computational entities which have elements that provide them with the ability to perceive and manipulate their environment: sensors and actuators. These are characterized by displaying various properties that allow them to adapt in order to achieve their objective. Autonomy, learning, collaboration and reasoning are examples of these properties, which together make them intelligent artificial entities. This article shows the development of a framework that has made it possible to speed up the construction of a system of adaptive mobile intelligent agents, called SySAge. The system agents have integrated search and learning techniques for the execution of automated processes focused on solving search, classification and optimization problems. It has been found that through learning, the agents were able to estimate input parameters and apply them in estimating the shortest route in a graph, considering cost and penalty aspects. To determine the choice of search technique, a probabilistic selection was used. The autonomous behavior of each agent was appreciated through the various attempts to solve the search problem and not to focus the information acquired individually on a single agent. The agents displayed a conditional behavior, depending on the experience acquired in previous executions, with which they were able to decide which search technique should be used to avoid incurring any penalty. The contribution of the study was mobility, which constituted an additional characteristic, incorporating in the agents the ability to move around and use resources from various nodes in a distributed environment. It has been concluded that if collaboration mechanisms are integrated, the agents could be able to share the acquired experience. If the automated learning mechanisms are used, it will be possible to remember the values acquired in past experiences (previous iterations) with which it will be possible to try to solve function optimization problemswell-prepared abstract enables the reader to identify the basic content of a document quickly and accurately, to determine its relevance to their interests, and thus to decide whether to read the document in its entirety. .