The paper presents a scheme for assessment selection of actions which the author proposed in an architecture meant for autonomous intelligent agents. A variant of this scheme was implemented for a relatively simple application, the components of the basic assessment being trained with 'Candidate-Elimination' algorithm and the components of the final assessment (prognoses-selection) with the 'Back-Propagation' algorithm. Some observations about the specific elements of this implementation as well as some preliminary results are presented.
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).