ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Luka Debeljak and Matevž Gačnik (2000) . Prešeren awards for students.

Full text not available from this repository.


Planning and plan-optimization of production processes are described within this work. Optimization approaches that are needed in modern companies are included in the developed system. A two-phase approach was used in optimization of production (manufacturing) planning. Parameters of production plan wew optimized during the first phase; second phase constructed the optimal production plan. Artificial intelligence, especially genetic and machine learning algorithms were used in optimization methods. A method that shows a possible way of teaching and algorithm to choose a set of optimal parameteres that enter the production plan is described. It is also shown how that selectionof parameters implies optimization and speeds up the planning process. A new production planning approach is presented using this optimization method. Process scheduling is a problem present in many areas of computer science. This problem becomes particularly intresting when processes are subject to a dynamic enviroment , i.e. changing conditions, demands and restraints. Typical goals for process scheduling algorithms are to minimize time, cost and resource occupation. More complex optimization tasks look at combinations of these criteria. The idea is too collectively improve the process in multiple fields. Production plan can easily be optimized using this methods when it is described as a process with its restraints. In this work, we also propose a general model for a genetic algorithm for process scheduling that successfully attacks the problem. Representations of solutions within the genetic algorithm are thoroughly described, as are the genetic operators participating in this evolutionary optimization approach and communication with preoptimization part that is done by machine learning algorithms. The result is an efficient two-phase intelligent system that successfully attacks the problem of production optimization. The system was also tested by integration into an existing information system, which was used as a realistic data source.

Item Type: Thesis (Prešeren awards for students)
Number of Pages: 111
Language of Content: Slovenian
Mentor / Comentors:
Name and SurnameIDFunction
prof. dr. Saša Divjak233Mentor
Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=51012&select=(ID=3482452)
Institution: University of Ljubljana
Department: Faculty of Computer and Information Science
Item ID: 3763
Date Deposited: 25 Jan 2017 09:49
Last Modified: 13 Feb 2017 14:18
URI: http://eprints.fri.uni-lj.si/id/eprint/3763

Actions (login required)

View Item