AIMS:
The aims are to develop student ability of modeling and analyzing real systems by means of discrete simulation, when applying the model, analysis of simulation results and comparison of alternative solutions. The student gets acquainted with artificial intelligence through models, structure of intelligence agents and machine learning. Knowledge needed for the application of artificial neural networks is acquired by means of simulation and software tools.
LEARNING OUTCOMES:
The student should be able to independently use software for modeling and analysis of real systems by means of discrete simulation, to choose methods based on the application of neural networks to solve engineering problems, along with analysis and presentation of obtained results, and to develop ability of team work.
THEORETICAL TEACHING (Syllabus):
Introduction to simulation. What is simulation, when to use it, terminology, model classification, types of simulation, steps in a simulation study, advantages/disadvantages of simulation study. The concept of discrete simulation, events list analysis. Simulation language GPSS. The application of simulation. Verification and evaluation of simulation models, output data analysis, comparison of the system alternative projects. Simulation of manufacturing systems. Artificial intelligence – definitions, basic concepts and paradigms. Knowledge databases, knowledge acquisition, learning models, decision tree, soft-computing development, autonomous systems. The structure of artificial neural network (ANN), neuron - processing element, activation functions. ANN models, learning algorithms, uncertainty system, non-linearity, estimation, clustering. The application of ANN.
PRACTICAL TEACHING (Syllabus):
General principles and examples of simulation. One-dimensional channel simulation, events management. Getting acquainted with software used for modeling and analysis of real systems by means of discrete simulation (laboratory work). Intelligent agents as a basis for intelligent systems development. Artificial neural networks in intelligent systems. Getting acquainted with software for artificial neural network simulation (laboratory work). Recognition systems, mobile robots, simulation of artificial neural networks (examples). Home works and seminar works related to simulation of real systems and application of artificial neural networks (recognition systems-computer vision; pattern recognition of standard manufacturing features of workpieces; recognition of gripping objects – robot vision system).
LEARNING RESOURCES:
[1] B. Babić, FLEXY-INTELLIGENT EXPERT SYSTEM FOR FTS DESIGN, Series ITS, Vol. 5, FME, 1994, 18.1 /In Serbian/ [2] Z. Miljković, SYSTEMS OF ARTIFICIAL NEURAL NETWORKS IN MANUFACTURING TECHNOLOGIES, Series ITS, Vol. 8, FME, 2003, 18.1 /In Serbian/ [3] Z. Miljković, INSTRUCTIONS FOR SOFTWARE FOR SIMULATION OF NEURAL NETWORKS-BPnet; ART Simulator,
http://cent.mas.bg.ac.rs/nastava/ksvi/index_ksvi.htm2007, 18.3 /In Serbian/