Authors: Smail Tigani; Mouhamed Ouzzif; Abderrahim Hasbi
Addresses: National High School of Electricity and Mechanics, Hassan II AIN-CHOCK University, Casablanca, Morocco ' National High School of Electricity and Mechanics, Hassan II AIN-CHOCK University, Casablanca, Morocco ' Mohammadia School of Engineering, Mouhamed 5 Agdal University, Rabat, Morocco
Abstract: The aim of this work is the improvement of cognitive agents performance. An agent is designed to follow fixed instructions to reach a given goal, this can be considered a limitation of agent technology because it does not have a minimum level of intelligence. This work proposes a new algorithm able to make prediction and learn from its experience in the prediction of a supervised environment. This allows the agent to analyse the history observations and make prediction of future environment state using the designed auto-adaptive algorithm based on stochastic models. The algorithms designed in this work can be applied in optimised scheduling or random environments management.
Keywords: agent technology; distributed systems; prediction algorithms; learning patterns; random systems; Markov chains; linear modelling; cognitive agents; stochastic modelling; multi-agent systems; MAS; agent-based systems.
International Journal of Intelligent Systems Design and Computing, 2017 Vol.1 No.1/2, pp.28 - 42
Received: 29 Jan 2014
Accepted: 30 Mar 2014
Published online: 10 Mar 2017 *